Seeking Alpha
We cover over 5K calls/quarter
Profile| Send Message|
( followers)  

Akamai Technologies, Inc. (NASDAQ:AKAM)

December 08, 2011 8:00 am ET

Executives

Dana Burd -

Adam Greenbaum -

F. Thomson Leighton - Co-Founder, Chief Scientist and Director

Natalie Temple -

Robert W. Hughes - Executive Vice President of Global Sales, Services & Marketing

Noah Fischbach -

Unknown Executive -

Robert Blumofe - Senior Vice President of Networks and Operations

Pedro Santos -

J. D. Sherman - Chief Financial Officer, Principal Accounting Officer and Senior Vice President

Paul L. Sagan - Chief Executive Officer, President and Executive Director

Analysts

Unknown Analyst

Natalie Temple

Hi, good morning, everyone. Welcome to our Annual Investor Summit. I'm Natalie Temple, Manager of Investor Relations here. I want to thank you all for coming today, and welcome those of you tuning in on the webcast that's being broadcast to you live over the Akamai HD Network. We're very excited about our program today. But before we get into our prepared remarks, we're going to go through the Safe Harbor statement.

This presentation and webcast include information about Akamai's future expectations, plans and prospects that constitute forward-looking statements for purposes of the Safe Harbor provision under the Private Securities Litigation Reform Act of 1995, including estimates about future revenues and market developments. Actual results may differ materially from those indicated by these forward-looking statements as a result of various important factors on the slide before you and in our Qs and Ks. In addition, any forward-looking statements represent our estimates only as of December 8, 2011, and should not be relied upon as representing our estimates as of any subsequent date. While we may elect to update forward-looking statements at some point in the future, we specifically disclaim any obligation to do so even if our estimates change. Please also see the note on non-GAAP financials. With that said, we're going to run a short video clip and then Akamai's President and CEO, Paul Sagan, will take the stage.

[Presentation]

Paul L. Sagan

Good morning. Good morning. Thank you all for joining us here in Boston. Today, that video should give you a sense of how excited we are about the opportunities ahead of us and how we help our customers move their businesses forward, faster. We're grateful that you came to Boston this morning. We've had some dicey luck over the years. A few of you will remember with Boston weather in December, so those earmuffs you were given come from all the years of telling us how cold you were. Those are not earplugs, you are supposed to listen to us this morning, and then we hope you wear them when it gets colder pretty soon.

The world is definitely changing or perhaps better said, it has already changed significantly since Akamai set out to solve the challenges of doing business on the Internet nearly 15 years ago. Looking back to 1998 when we started, only 4% of the world's population, less than 150 million people were connected to the Internet. For most of our customers, an online presence consisted of little more than a text-based brochure-ware website. Today, the number of people connected to the Internet has skyrocketed to more than 2 billion, with exponential increases in their connection speeds. If the number of people connected has increased, so has the complexity of our customers' websites to be highly dynamic online properties, full of rich media. Predictions for 2020 put the number of connected people on the Internet at more than 5 billion, with more than half of those at broadband speeds. Improved accessibility and increasing speed help to change the way businesses, our customers thought about how they could use the Internet to their advantage.

From those early days of brochure-ware sites to today's world of highly complex online properties, our customers have come to rely on Akamai and Akamai innovation to help them stay ahead of all this change. And in the face of this continued evolution, we don't believe that, that reliance will change any time soon. Our customers today are faced not only with exciting opportunities, but also huge new hurdles to operating their online businesses. It's increasingly clear that we are no longer living in a world where people are simply connected to work, to friends, to family, to content and applications, but rather, we've all become hyper-connected. We're operating in a world where consumers and workers no longer go online, working, shopping, playing, communicating, being entertained. The pause button for our customers has completely ceased to exist. At the same time, we're using every kind of device imaginable: computers, smart phones, tablets, connected TVs to stay connected in our homes, our offices, our cars and even now, up in the air. Our industry hasn't finished coming up with exciting, new ways to keep us connected, and Akamai has been and will be there each step of the way. We're providing critical cloud infrastructure and content delivery services that help our customers across many industries from media to manufacturing, from healthcare to financial services, from the public sector to private clouds, accelerate through innovation and capture the incredible business opportunities that are ahead of them.

It really doesn't matter the content or the type of activity or the application. We all want instant access, and we want it anywhere from any device. And let's not forget that we demand a consistent and a safe experience optimized across all these devices that we're using on a daily basis. It's a tall order, but it's the one on which we are already delivering for Akamai customers, all the way to their end users. This state of what we call hyper-connectivity was certainly evident over the Black Friday, Cyber Monday weekend of the holiday shopping season. At its peak, traffic across our platform to commerce sites reach nearly 2 million page views a minute. Perhaps even more exciting were some of the other correlations that we saw in some of the data. Traffic to social networking sites peaked at exactly the same time as traffic to commerce sites, seems to imply that consumers were using these online social channels to uncover and share opportunities for savings.

Further at the peak on Cyber Monday, 20% of retail traffic on our network versus traffic to retail sites was coming from mobile devices, as consumers appear to be taking advantage of last-minute online deals while they were out and about. We've never seen anything quite like that. Through ongoing innovation and development in the Akamai platform and expanding ecosystem of partners and even important changes in the ways we operate our own business now, we continue to address the evolving needs of our customers as they capitalize on opportunities like these presented in this hyper-connected world. Any company that wants new business in an environment needs to be prepared or be prepared to be left behind.

Over the course of today's agenda, you'll hear from Tom Leighton, our Chief Scientist and Cofounder. Tom will take you through our vision for the services and technology that we're developing in Akamai labs to give our customers the solutions that address today's business needs and tomorrow's. Tom's got some great live demos planned for you that's going to showcase the Akamai innovation engine. I think it's going to be the highlight of today for you. Bobby Blumofe, the Senior Vice President who runs our platform will talk to you about how we're working to continue expanding the reach of the Akamai platform, bringing its intelligence into places it's never been before, including into cellular networks. He'll also update you on the evolution of our network relationships, and how this is driving new business models and revenue opportunities for existing network partners and new network partners in the future and for us.

I mentioned earlier that we continue to evolve our own business to best meet the changing environments and requirements of our customers. Increasingly, they're looking to Akamai to provide a total solution, both products and services to support their cloud-based initiatives. Bob Hughes, our Executive Vice President of Global Sales, Services and Marketing, will provide you some more detail on our evolving go-to market strategy. He'll be highlighting our full suite of solutions on the Akamai intelligent platform, as well as the expanding global reach of our sales efforts designed to bring Akamai closer to our customers. And you can hear some specific customer examples in how organizations like Harley-Davidson, the NFL, Hertz and Arabian Beer [ph], using our solutions to innovate and transform their businesses online. Bob is also going to talk about the value of our channel and our technology partnerships and how we think we can leverage them to drive even more growth going forward. And then to close out, J.D. Sherman will provide a review of the financials and what we believe is an extremely attractive and robust business model.

Finally, you may recall that in the past few months, we've announced 2 new senior appointments at Akamai to the management team. Both Rick McConnell, EVP of Products and Development and Kumud Kalia, our new CIO, are here today, so I encourage you either during the breaks or at lunch to visit with them and get to know some of the newest people helping me to drive the company.

Before I turn the stage over to Tom this morning, I'd like to share with you some thoughts about why I'm so excited. I think why everyone at Akamai is so excited right now about our future even after more than a decade in business, and why I believe that we can continue to drive growth in our business and continue on the target that we set to grow our business fivefold to $5 billion in this decade. Of course, that's going to require that we grow the business at a compound annual rate of at least 15% over this period. This is a big challenge. We recognize that. But given the markets that we're in and the foundation that we've set across the globe and on the Internet over the first decade, we are extremely confident that we're going to do this. We're seeing growth in the industry segments where we operate, growth on which we believe we have great potential to capitalize. For example, Gartner forecasted spending in the app acceleration market will grow at least 13% through 2013. They also predict that the market for cloud, public cloud services, which grew 27% from '09 to '10, will be worth more than $177 billion by 2015. This represents a 5-year compound growth rate of 20%. We believe that we can grow our enterprise business at least that fast.

In addition, IDC forecasts global consumer spending on e-commerce will go from $700 billion last year to nearly $1.3 trillion in 2015. That's an annual growth rate of nearly 13%, and we believe that our commerce business can grow at least this fast or faster. Finally, IDC forecasts that the monetization of online video, whether it's through advertising, subscription, pay-per-view and other models, could grow at 40% compound over the next few years. Industry estimates still have the amount of media consumed in the home over IP or over the Internet at just 1% or 2% of all viewing. We believe that, that will expand dramatically, that the data loads caused by online video could grow by over 100 times in the next 10 years, which means that even in markets like video delivery with consistent unit price compression, we believe there's massive opportunity to capture this rising level of volumes and grow our media business for the second straight decade.

It is not just market predictions from the industry analysts and observers that drive my confidence in our ability to meet our growth objective. We realized so much of that potential already in our business. As you'll hear later from J.D., our cloud infrastructure services business has gone grown at approximately 20% year-over-year-over-year. And even despite periods when our content delivery business growth has been somewhat muted in comparison, we believe this content delivery business has a potential to continue to achieve close to 20% annual growth rates or better because of this flood of demand that's coming.

We believe strongly that this hyper-connected world accelerated by the trends of cloud computing, pervasive mobile connectivity and Internet video is going to continue to disrupt existing business models. This is going to drive more and more investment by enterprises into online initiatives that will increasingly demand better performance, scale and of course, security. And we think that will open the door for the introduction of more and more new services, including some of the things that you're going to see demonstrated this morning. We recognize these opportunities and we're committed to providing to our customers the right technology solutions to take advantage of these opportunities. The emergence of cloud computing, especially hybrid clouds, is a prime example of how businesses of all types are adapting to this hyper-connected world. Cloud offers on-demand access to content and applications from anywhere, anytime, as well as cost and environmental business for businesses.

Yet moving to the cloud creates new technology challenge that many companies rightly have doubts and perceive the cloud as having some serious drawbacks, particularly in the form of performance issues and security threats. We try to remove much of this complexity of moving to the cloud and address the needs for pervasive secure connectivity to applications hosted in a variety of different environments. I talked about our expanding partner ecosystem, and cloud computing is one of the areas of this ecosystem that we think is coming together really, really well. We're working with cloud providers to make it easier for them to offer optimized capabilities to all of their customers. We're bringing to market exciting new technology that allows them to bring Akamai intelligence behind the firewall into their networks and data centers, places that it hasn't been traditionally. This allows what we like to think of as Akamai's applications to flow from the data center to the cloud all the way to the end user.

Mobility and hyper connectivity go hand-in-hand, and advances in mobile technology are changing the expectations of consumers and business users alike. Demand for instant access to dynamic and personalized content and applications is now the norm. Akamai is responding with current and future solutions that not only support device-specific optimization, but through our partnership with Ericsson, the ability to extend the Akamai intelligent platform inside cellular carrier networks. In the world of media and entertainment, video consumption is happening on a myriad of devices like these and hundreds others, and so it has to be available on the go. And as more and more video moves to the Internet, it's being delivered with greater quality, TV-like quality and more often, it's being delivered over Akamai's HD Network. At the same time, we're giving our media customers the tools to look closely at what the end-user experiences to understand how their viewers are responding to the content that they're consuming or that they're being offered, giving our clients insight to help them expand their audience and generate new revenue streams.

Despite the promise and the excitement of today's hyper-connected world, we're aware that there are big risks that it poses. For their customers and their users, the online world can be a very dangerous place to do business. Just last month, we've tracked 1.5x more distributed denial of service attacks than we saw in all of 2009. Not only is the number of attacks increasing, but attackers are more sophisticated, and their attacks are becoming more complex and harder to stop. The damage to a business, even just by the loss of reputation that these attacks can cause, is enormous. To help mitigate these risks, Akamai offers a range of security solutions that can help to take the worry out of doing business online and in the cloud, and that allows our customers to try to concentrate on doing what they do best, driving their growth in innovation.

We recently hosted more than 500 customers from 30 different countries at our Annual User Conference here in Boston. We've heard directly from them about what we're doing well and where we can improve, and how they're using our technology to support the growth of their businesses. But not only that, we demonstrated our dedication to powering the acceleration of innovation in our hyper-connected world now and in the future. Make no mistake. Akamai is committed to staying at the forefront of Internet innovation. We have programs in place to make sure that we're hiring and retaining world-class talent at all levels of our organization. We're thinking differently about our ecosystem of partners and how we can work better together to solve our customers' most challenging business needs. We're working hard to create for Akamai, Akamai customers for life, using our technology across all aspects of their online businesses. And we're dedicated to being a truly global company, opening new markets and serving our customers in those markets locally, all of which translates into driving positive returns for our shareholders.

Thank you again for joining us this morning. We have a full agenda, and I'm sure you're going to find it very, very informative. And now before I turn the podium over to Tom and his demos, I'd like to share with you a short video from that Customer Conference. I think it will give you a good sense of the excitement that we and our customers felt at that event. Thank you.

[Presentation]

Unknown Executive

Good morning, ladies and gentlemen. Please join me in welcoming Akamai Cofounder and Chief Scientist, Dr. Tom Leighton.

F. Thomson Leighton

Thanks very much for joining us today. As I said at the Akamai Customer Conference, we are now living in a hyper-connected world and it is true that in the hyper-connected world, everyone and everything that can be connected to the Internet is connected to the Internet. Here are some examples. The first example, that's not my 15-year-old son playing video games, but sometimes I wonder. He spends pretty much all of his time online. He owns all sorts of connected devices, accesses all sorts of entertainment and communications through the cloud, and likes to have access moved from one device to another seamlessly. In fact if I want to talk to him, I do it through texting on my phone even if he's down the hall. He really is connected in the new hyper-connected world.

Shoppers are more and more online. In fact in mobile devices, as Paul noted, that 20% of online purchasing now is being done on mobile devices. And in fact, you use a mobile device even if you're in the brick-and-mortar store or trying to get to the brick-and-mortar store, and that site better be fast, not only on the landline device but on the mobile device, and we'll talk a lot about that this morning.

Also, the IT professional now needs quality of service for the mobile workforce, has to move this infrastructure into the cloud and is really scared to death about the security challenges and for good reason, as we'll talk about this morning. It wasn't always this complicated. You go back 10 years ago and putting up a website was pretty easy. You bought some web servers, stuck them in the data center and hooked up your transit pipe and you were online. Now, over time, traffic grew and you need to buy more web servers and some load balancers to balance the traffic, then you needed to make the site do something, not just have a static banner up that maybe run some applications. That meant you have to buy application servers and manage them. Then you needed to restrict the access to certain information if it was confidential, and you needed to support sales online, so you had to have SSL or run HTTPS, so you could have secure transactions to your application. Then you had to make the site be sexy, so you had to have streaming video. You had to be -- make it rich with interesting-looking content, and that meant more servers to manage the video.

Traffic grew again. You needed more of these devices and load balancers, and then a weird thing happened. You had so much traffic and all these servers were so expensive, people started developing servers for things like compression and connection management and forward-caching that would offload the servers you already bought. So you're buying more servers to offload the ones you've got, so it's getting much more complicated to manage the web presence efficiently. Then bad guys came and now you needed intrusion detection systems and intrusion prevention systems to manage that. Then the protocol started changing. We're moving from IPv4 to IPv6. That's a big overhaul. You want to be able to support DNSSEC. Now in this country, most folks don't care about DNSSEC. We support it today. We probably have a dozen customers using it, but it's something that everybody needs to be thinking about in terms of putting security at the DNS layer.

When you're all done with that, you got a big data center and now your Risk Manager says, "You got to build that data center in multiple geographic locations, so if something bad happens in one place, we're still online in another place." So you replicate the whole thing and even then, you're not done because now you're just beginning to think about the network, the Internet.

Now the good news is the Internet, I believe, is the invention of our generation. Maybe you can say there's going to be stuff in genetics and computational biology that will change medicine and the way we live, and that could well be the next major invention. But in our generation, it's changing everything, the way we work, the way we play, the way we learn, the way we communicate. And the expectations of the quality of the Internet have changed dramatically. Ten years ago, if you clicked on the URL, it took 10 seconds to load, and you had some broken objects, okay, that's the Internet. That's the way it works or doesn't work. Today, people have a different expectation because now, we all are critically reliant on it. We expect that page to paint in 2 seconds. If you're running one of the major search engines, you need 1-second response time globally, okay, or somebody's in big trouble. It has to be that fast. And if you're watching video, you expect it to be in high quality and if you change the video, it's like changing channels on the TV, it needs to change right away. That's the expectation. The Internet needs to be like TV. The problem is that it's not, and it's not constructed like the TV broadcast networks. In fact, it's not one network, its 12,000 networks. A lot of them in financial distress, and a lot of them not getting along with each other.

Today, when you take all those boxes, sometimes, you'll still put them in your own data center, a lot of that still going on. Sometimes, you put them in a cloud data center, which really means you're hosting it somewhere else, maybe on a set of VM servers that you got. The problem, either way, whether in a cloud data center or your own data center is typically, you're far away from your consumer, your end user. And as those bits work their way through multiple networks, bad things can happen. It can be slow and you can have poor quality, resulting in a subpar experience for those users who aren't right next to the data center, which most of them aren't. You have a particular vulnerability at the origin, at the data center. This is where you're most congested and you see this during the holiday shopping season when a lot of people go to one e-commerce site to take advantage of the sale. That becomes the bottleneck area. It also becomes a critical point of failure when bad guys want to bring you down, and we'll give lots of examples of that today. They just go attack the data center, flood your first mile with traffic, and you're down. There's also continuing problems between the networks as they get into economic disputes such as the occasion when level 3 de-peered with about 2 dozen other major providers. And that meant that anybody who is connecting through one of those providers, say, Cogent, could not reach any website that was hosted through a level-3 data center, and that lasted for a couple of days, unless, of course, they were Akamaized and we can just see how that can work. For example, say, this kind of -- I'll get to there in a minute, so these are the problems that led us to create Akamai, why we're in business, to make the Internet work better. And at the high level, the way we do that is we take all the functionality that I talked about, it's an all those servers and we put that functionality in software in a single Akamai server, a piece of commodity hardware that's fabricated just for us, and then we make lots of those servers and we put them everywhere we can. We spread them all around the Internet in thousands of locations. And so then, instead of the end-user have to going across all those networks and across the country to get the cloud data center, very often, they can get the content locally from an Akamai server right across the last mile, in their network, in their city.

If the origin server is down or the cloud data center is down, the end-user still can get access to the site. They may not be able to complete the order, but they can maybe file an order, and we'll deliver the order later. In fact, recently, there was a very major cloud provider that had a major outage for many hours during the day, and those -- the customers that use that major cloud provider, those that were Akamaized were fine, no one knew they were down. In fact, people didn't realize the magnitude of the outage of the cloud data center because many high-profile users of that cloud provider were using Akamai, and we present to the site pretty much as if the site were operating normally.

Now in the situation I was talking about when a major network de-peers with other networks or there's a disruption of some kind or poor performance in the Internet-given links and we need to get to a database that's only in the data center, we developed our own routing protocols that sit on top of the Akamai servers, our own communication protocols and our own application layer protocols. In effect, we've built our own Internet, our own virtual Internet, and so we can route around that problem and get access to the data center, so we can complete the transaction. So even for non-cacheable content, you're doing a transaction. Just for you, we will make that transaction be a lot more reliable and a lot faster because it's using the Akamai Internet, which sits on top of the real Internet.

Now today, we do this on a platform in 1,900 locations inside 1,000 networks. I'm not talking about peering. We peer with a ton of networks. We are physically inside 1,000 networks. There is no entity like Akamai in this business model today. We're the only ones that do this. We're in hundreds of cities in 75 countries. Basically, we want Akamai servers wherever end-users are, in their city, in their network and deep in the network close to the end-user because that's how you get great performance.

Today, we deliver over 150,000 websites using this platform, including 90 of the top 100 retailers, 49 of the top 50, all the top media and entertainment companies and a majority of the top global banks, and we work with all of the major Internet portals. There's a lot of traffic riding on the Akamai platform today. Typically, the daily peaks are over 5 terabits a second, when you saw Mike Afergan in the video talking about our record traffic day that was over 8 terabits a second of traffic. We deliver over 20 million items every second, way over 1 trillion deliveries per day, and a delivery is when one of you or somebody out there has a browser and you go ask us for something, you don't know you're asking us, you're asking the website that we're hosting or carrying for it, we deliver to you and log it and we're doing that over 20 million times a second.

Let me show you this slide, what's going on right now in the Akamai platform, so if we could switch to GNET over here. This is actually a live view of what's going on in the Akamai platform. Right now, we're delivering close to 4 terabits a second of traffic and over 16 million deliveries per second. Now we're in sort of a high time of year in terms of deliveries per second because there's a lot of e-commerce buying going on, but we're early in the morning on the East Coast. So the West Coast isn't up yet and in Asia, it's getting later in the evening. So the peak traffic will be higher later in the afternoon on the East Coast. We have servers in any location where you see a light. Any city with a light is where Akamai servers are located. And if there's a big spike coming out of that city, it means we got a lot of servers there with a lot of connectivity. If it's a little spike, not so many servers, maybe as few as 2 and maybe it's on a shared ethernet. If the spike is yellow, it means we're using that capacity a lot right now. If it's blue, we're not. Green's in between.

Now I can stop this globe and zoom in on any city on the world that you want, and we can see all the locations where the servers are in that city, and then I can pick any of those locations, zoom in on that and we can see every single server, what it's doing. So where would you like to go?

Unknown Attendee

Europe.

F. Thomson Leighton

Europe. What city in Europe?

Unknown Attendee

Barcelona.

F. Thomson Leighton

Barcelona, okay, let's go over there. All right, let's see, I don't know if we're in Barcelona.

Unknown Attendee

Prague?

F. Thomson Leighton

Prague? Probably in Prague, but that's -- well, we're in Spain. We're in a lot of places in Madrid, the Madrid area. So every sphere there represents a location. So not only -- most companies, you think of having a data center in Southern Europe maybe, probably not in Spain. In Madrid, we're in several dozen locations and the codes there, the acronyms are the name of the provider whose location we're in. And you can see they're all in Madrid, the second acronym. So you can see T-Systems. I don't recognize EPHH [ph], Globix, France Telecom, Tiscali, Verio. So let's pick one of them here. Okay, so I picked one of them. This is Tiscali, and you can see every line here is a server, starting with its IP address and these have the S-code. That means they're running SSL. So this is a special set of servers that's doing HTTPS traffic. They're probably doing credit card transactions for the holiday buying, for example. They would hold our customers' search for e-commerce, for banking, for any kind of activity, where you want to have it be confidential and you can see how many hits a second each one is doing. So anywhere from 13 to 186 deliveries a second, then you can see the megabits a second, overall probably doing 50 to 100 megabits a second. And you can see it's very lightly loaded right now, probably about 10% load. And then you get a guy here with an L, an L stands for leader. He's an elected leader by the other servers in this cluster and in every cluster, there's a leader, distributor leader like Chanel [ph] then goes on and that server will resolve any disputes that arise in between the servers in this region.

All of the control is distributed and it's all localized and decentralized, and so there is a hierarchal leadership that's all distributed. If something were to have happened to cut this region in half, the backplane were cut in half, each happen to elect its leader. Disputes could be about where to send an end-user for content. There could be disputes about which servers are performing well. In this case, they're all up and fine, running just fine. In many regions, if we could go visit, you'll see down servers and maybe other servers have decided they didn't like how it's performing, maybe there's a disk failure. And then immediately, that server is taken out of commission and other servers take over its work.

All right, I'll try another one here. Let's see, here's France Telecom. This is running W-service, which is HTTP, and that's the majority of our traffic is HTTP. That would be any download you do. It would be all of our HD video service, would be our EdgeComputing service. Things that are not secure would be on this service, and it's got more load going here, about 30% to 40% load. And it’s running several hundred megabits a second, with thousands of deliveries a second, okay? Let's see, somebody said somewhere else that...

Unknown Attendee

Brazil.

F. Thomson Leighton

Brazil? All right. Let's go to Brazil. Yes, there's Prague. We could go back and visit there, but let's get down to Brazil. All right, so we have several locations in Brazil and everywhere again you see a sphere, there's Akamai servers located there. Let's see who this guy is. This is CTBC, who I don't know, and it's in Uberlândia, which I am not so good in geography. Also, I don't recognize, but there's a fair amount of traffic there. We have 9 servers doing HTTP service, all operating. Thousands of deliveries a second and hundreds of megabits a second, and pretty highly loaded here, about 70% load and we'll start doing load-balancing when it hits 100% by this measure, and 100% does not -- you're still, we're still fine at 100%. It's not like it's maxed out of CPU, but it's when we start balancing traffic among other regions and other servers. There's many other cities in Brazil. We have lots of locations and again, I don't know of many companies that run data centers, U.S. companies or European companies that have data centers and points of presence in Brazil, and we've got dozens there because there's a lot of end users in Brazil, its several percent of our traffic on a daily basis.

Okay, let's switch back over. So again, this is to give you an idea of how different we are. Some of you cover some of our competitors and maybe they're in a dozen locations, maybe 2 dozen, the guys that are most distributed. We're in thousands of locations, and you pick a country like Brazil that's very important. We're in dozens of different locations in Brazil, and that's one reason we can get such great quality and such great scalability on a global basis. So the benefits are pretty clear. It's a lot faster using Akamai than using anything else and if you're doing video, it's a lot better quality because that video is coming across the last mile, and you get much higher throughput which translates into much higher quality. It's much more reliable. There's no data center to go down that cause problems. At any given time, we didn't see anything. I don't know that happen, but we have thousands of servers that aren't working. We just didn't see any on display on those 3 places we went. There's no single point of failure. There's huge capacity and scalability, and it's all on-demand. We didn't know that day, that the customer conference was going our biggest day ever that we were going to spike several terabits a second of extra traffic, but it was fine.

It becomes very cost effective for our customers because they don't have to worry about provisioning for the holiday season for all that traffic. Okay, their back-end applications, they worry about that, but the front-end presence, we're taking care of all of that. Often cases, 95% to 99% of their load comes on to us, and they pay just for what they use. And then we'll talk a lot today about security and why this cloud presence, the true cloud that Akamai enables makes it so much more secure for our customers.

So looking at the growth over the last 5 years, it's been pretty amazing, and I wanted to share some statistics there. One of them is the number of unique IP addresses connecting to Akamai, which has about tripled over the last 5 years, and these are the folks we see on a daily basis. The content has exploded to the point where now we serve the equivalent of 10,000 volumes of the Encyclopedia Britannica every minute, and that's not just the print version, that's the online version with pictures and videos, and DVDs, all that put together 10,000 every minute is equivalent of what we're doing in traffic. It's really pretty large scale.

Now if you look at mobile, as Paul said, mobile's exploding, and we don't have exact measurements on where it was in 2006, but it was less than 3 terabytes in a day and much bigger and growing very quickly. Commerce is growing rapidly, online and through Akamai and now, on a daily basis, we handle more than $0.5 billion of e-commerce. The bad news is attacks have also grown even faster in some cases. So that now, the attack sizes are measured in the hundreds of gigabits a second for the largest attacks.

Now I think it's interesting to look at -- okay, take the last 5 years, a lot of growth. What would happen if we took the same rates of growth in each category for the next 5 years and look at that, and do we think that's reasonable or not? So if you took the same rate of growth and the number of IP addresses we see in a day and project it forward, that would be 1.6 billion we've seen a day. Now my guess is we're going to see a lot more than that because more than 5 years, IPv6 will be pretty prevalent, and we're going to be seeing zillions of IP addresses. It's -- IPv6 is going to change the infrastructure for how a lot of things work.

Content growth, that would be 1,500 petabytes per day delivered. And that's a whopper number, I think probably small compared to where we'll be as you think about so little of video is online today. I think there's a pretty decent chance that we really see that change over the next 5 years. I don't know exactly when it will happen or what the trigger will be. But if we're only doing 1% or 2% today and a relatively low quality, imagine the levels if you got 20% or 30% at high quality. It'll grow much faster in that case.

Mobile, for sure, this number is conservative. Mobile will take a higher and higher share already, and you can see the share now on a bits basis. It's a small fraction of the total bits today, and part of that's because of video and software downloads. Commerce transactions are already a nontrivial portion of 20% today for page views on commerce. I think the mobile number will be much larger percentage, much larger than you see here. Commerce, very reasonable to think that commerce in the platform will more than triple in the next 3 -- the next 5 years. And unfortunately, I think the attack traffic, again, this will be a conservative number, we'll be seeing attacks in many terabits a second within the next 5 years. So we think about how fast it's been growing. I think it's probably conservative compared to where it could be over the next 5 years because we think of the Internet as being so pervasive, but it's still just at the beginning. A lot of potential. So that's what we're all about and we're just trying to enable this potential, make it real, make the Internet work better for our customers. And I'm going to spend the rest of my time talking about the 4 areas that we're focusing on: media, performance, security and cloud, and give examples of our products and new technology in each of these areas.

And we'll start with media.

Video quality has come a long way. And I want to show you 2 examples. The first is the typical example that you're probably familiar with of a video online. So if we could roll the first clip.

[Presentation]

Okay, now you've all seen lots of videos like this. I find this a pretty fun one, entertaining. You noticed the quality wasn't that great, but it didn't need to be. You're watching it for 30 seconds. Hard to imagine spending 2 or 3 hours leaning back in the living room watching that. You're tired of it after a while, the quality is not so good. But quality is changing, and now we can deliver a very high quality video on the Internet. And maybe we could see the next clip?

[Presentation]

Okay now, you can, and lots of people now do, watch that for hours at a time online. Because the quality is much better, it's premium content and that kind of content now is available in high quality online. So it's really changing what's happening with video.

Moving to the Internet. Now it's not just sitting back in your living room. Media today online is about any experience on any device anywhere. And there's lots of places people want to watch high-quality video online today.

Now there's challenges in making that be a reality, and several. First, you've got to get high quality on any device anywhere, and there's lots of different devices out there and they run lots of different formats. And so it's not a matter of just streaming to the device, you've got to figure out what format the device has and make that stream available in all the different formats. The different devices are in different places and they have different bit rates that they can support. Now the higher bit rate, the better the quality, so you always want the most you can have. But in many cases, you can't get very much. And so when you're delivering the stream, you've got to be able to deliver it in multiple different bit rates so that any end user can get the best possible quality at any time.

You've got to manage the cost of all of this, and it's getting increasingly expensive to do it, especially as you go to the high-quality bit rates because that's a lot of bits and that means more expense and infrastructure. And of course, the media guys have got to monetize it, and a lot of them are just barely getting to the point where they can make more money than they're spending on the infrastructure and that's a continuing challenge in the media vertical.

And they've got to worry about security, which is a real problem because people will -- if you put the high-quality content online, they'd want to steal it and make it available for free. And you've got to worry about rights management. Who gets to see what on what device, when, and who gets paid for it? And that's a whole complicated area right now that really is turning the media industry upside down. Who's going to be on top? How does the money flow and who takes what cut?

Now Akamai can help solve all these problems. First, of course, we're always about quality and performance and the HD network is built to deliver high quality to any device anywhere. We will re-encode, transcode the content, the stream, into the varying formats so our customer can give it to us in one format, say, RTMP, if they're running legacy Adobe, and we can make it work on the iPhone and any other device, the iPad. Because we'll re-encode it in our network to deliver to the right format to the device. We also now, in the lab, can re-encode at all bit rates so the customer only has to give us the highest bit rate they want, we'll re-encode at lower bit rates to work for any device so that every device gets the best possible encoding at any given time. And it can change over time for the same device, as the end user moves around or traffic gets on the network.

We're also investing a lot in reducing cost. Now, you all been, and we too think about that every year, the price per bit to buy media in Akamai declines, and by a lot every year, has been that way since we have been in business. And there's the question of when is that going to stop? It's not. In fact, we're working hard to make the price come down even faster. Because by doing that, that enables so much more content to move online. And just some of the things that we're investing in to lower cost, our work on hardware and softer efficiency. So the same box that we've got today can deliver twice as many videos next year, okay? Or twice as many bits per second. And that makes it much more efficient. The co low cost for that box, the depreciation on that box is used to deliver a lot more traffic, which generates more revenue. We've also developed client-assisted delivery. Now in the old unsafe world, you could think of this as peer-to-peer. It's not the old peer-to-peer. It's special technology that we've developed that's highly controlled and highly secure, that runs on your devices and we use to deliver from your devices to save cost that we can pass on to our customer. Now we've been doing this for some time with the major software downloaders. So probably many of you, might be all of you, have Akamai software on-board your devices. You probably don't think about it and don't know it. And in some cases, we use those devices to help us deliver content. In fact, just last week, we did our first live event with our client-assisted delivery with the NFL on the Thursday night game. And so any of you who watched that game, if you click on the box that said you wanted the HD version, you downloaded Akamai software and that software was used to assist us in delivering the video to other users. Now today if we could switch over to -- I need this, I think we're going to do it up here, get the camera here. I'm going to show you the spinning globe for our client network.

Today, and this -- if we could switch over to the mobile cam here? Today, we have about 30 million clients that assist us with delivery. And wherever you see one of those spikes or lights, it is one of 1,000 randomly selected cities where we have clients. Now there's actually a zillion cities with clients. So we took 1,000 clients at random, out of the 30 million, about 29 million, and we've displayed them here and lighted up a city if we had a client there. And I think the fun ones to go like a remote region, like say, we go into Russia, really in the middle of Russia where there's actually people there. All right, let's get a little closer. Right now, you can see where the cities are. And we'll go to Kansk. And we can see in this city, if you could zoom in on the text, I don't know if you're able to read that, we've got 10 clients in Kansk in central Russia, helping us deliver content. Let's go here, if we could focus in on this one. I don't think I can pronounce it. And there is 1,470 clients in that city, okay? So here you are in central Russia, we now have delivery capability from clients for medium software download content there, which certainly improves the quality and reduces our cost. And then we can pass that cost savings down to our customer. And our goal is ultimately that every device will be fully integrated into the Akamai platform and an extension of the Akamai Edge. We made a lot of progress there over the last year. Okay, thanks very much, if we could switch back now.

We're also exploring new technologies. For example, multicasting. And I'm all set with that, so we're okay on the mobile. Now with multicasting, it doesn't really work today. You can't use it from one network to another. The idea behind multicasting is sort of the media companies' dream. It's the old broadcast model. With the old model of TV, they send the signal up onto the satellite and everybody gets it for free after that. Satellite's pretty expensive, but after that, you don't have any marginal cost for a new user. And multicasting is a technology that's been around forever, but never effectively implemented or used in networks that have the same idea. Send it out once, the routers replicate the content.

And it's an area that we're exploring that can help with live content delivery, and again to reduce cost. And because we're in all these networks, we're in a good position to help enable that within our network partners. And as you'll hear a lot about later, when Bobby Blumofe comes up, we're working on deepening our network relationships and getting our equipment and our software deeper into networks to help decrease cost further, also improving performance.

So a lot of effort at Akamai you don't hear about a lot to decrease our cost, which helps our customers and helps our margins to do that.

We also work in trying to help our customers monetize their content through analytics tools, both for performance analytics and business analytics so our customers can see what end users are watching, how they're watching it, how they move around in the site, and that helps them monetize the content better.

Security is a big area for us. We'll talk a lot about it later. There's special issues with security around media because you don't want that high-quality video to be stolen and then distributed for free. And there's a lot of work to be done there. The standard ways of doing this are not very secure. It's really security by litigation, which is just not effective for people who are stealing it offshore, which they are today.

And finally, with rights management. And this goes to the core of what's happening in the media business today, the media vertical as to who gets paid and how does the money flow? Is the content guy on top? Is the network guy on top? Is Google on top? And they're trying to get that worked out, and some models are emerging, such as TV Everywhere and UltraViolet to handle the rights management. And to show you how that works, I'd like to do it live because we've done some pretty cool development in this area. And to help me is Adam Greenbaum. And Adam, can you come up and help explain what TV Everywhere is all about and what we do and how it works?

Adam Greenbaum

Sure. So TV Everywhere -- switch to Mac. TV Everywhere, we view as authenticated and authorized access to content. TV anywhere is the notion of just going over the top to devices, which is something else we can do. So...

F. Thomson Leighton

All right. So say I want to watch a video. I want to see the Big Buck Bunny.

Adam Greenbaum

Sure. So this is an example of what would be a site where you need to authenticate yourself before we can watch any of the content. And as the site says, you need to be logged in, so go ahead and sign in.

F. Thomson Leighton

All right. So, sign in. All right. Well, what's that?

Adam Greenbaum

So this gives you the opportunity to choose from a list of different people who can provide you identity. So in the notion of having a site which has cartoon content, you might be required to log in with your pay-TV provider. MVPD (sic) [MVPDx] is the new hip term for MSL, its multi-channel video provider distributor. It's a big mouthful, but basically it just means your pay-TV provider.

F. Thomson Leighton

I see. So I could use my Comcast or FiOS account there and sign in?

Adam Greenbaum

Yes.

F. Thomson Leighton

And/or even Facebook or Twitter, right?

Adam Greenbaum

Yes. So content providers could do deals with something like a Facebook or Twitter to enable different functionality, a higher bit rate, something else like that if they want to do that.

F. Thomson Leighton

All right. So I'll go into my provider here and log in.

Adam Greenbaum

It's supposed to be the equipment providing your Comcast or Verizon FiOS credentials.

F. Thomson Leighton

Okay.

Adam Greenbaum

Great. So basically, what happened was we've taken identity from somewhere else and provided that into this experience to prove that you're allowed to watch these pieces of content.

F. Thomson Leighton

Okay. And it knows you had news [ph] so I'm now going to start to -- I can watch the movie now. So I pick the one I want and I say go?

Adam Greenbaum

And you say go. We are using the hotel Wi-Fi?

F. Thomson Leighton

It's going.

Adam Greenbaum

As fast as we can get over hotel Wi-Fi.

F. Thomson Leighton

Is this coming off this machine? Or where is this...

Adam Greenbaum

This is coming off the hotel Wi-Fi over the Akamai HD Network.

F. Thomson Leighton

So it's coming out of Akamai now into the hotel. See at Akamai, we like it that when I click that thing, I want it up right away and it took one or 2 seconds. So that's why we want this to be really fast. We are doing it over the local Wi-Fi. So it wasn't too bad with a couple of seconds there.

Adam Greenbaum

Yes. This is being adaptively delivered over the HD Network in qualities from 500 kilobits a second up to about 3 megabits a second.

F. Thomson Leighton

Now that looks like pretty good quality. What do you think that one is there?

Adam Greenbaum

We're probably at the 2.5- or 3-megabit per second stream.

F. Thomson Leighton

Okay. Yes, I've got pretty good resolution there.

Adam Greenbaum

Yes.

F. Thomson Leighton

That's better than SDTV. An SDTV is about 1 to 2 Megs.

Adam Greenbaum

And we're on giant screens, too.

F. Thomson Leighton

Yes. This is Internet TV. This is coming off the Internet. This is live. It could stop at any minute. Hopefully, it won't. All right, well now that's pretty cool. But what does that have to do -- is there anything more I can do with this thing?

Adam Greenbaum

Sure. So let's say you wanted to -- say, you know what, I need to go in the car, go to the store and let my kids keep watching this cartoon content. So why don't we go ahead and pause the content?

F. Thomson Leighton

So I can transfer devices here?

Adam Greenbaum

Oh, yes. Pause the content.

F. Thomson Leighton

All right, so I paused. And let's see. So I paused at -- it says 1 minute and 17. Can you all see that? Because that's going to be important here. All right, I'm going to switch this now to the iPad. This is a good example. My kids are watching this and I've got to go to the store and drag them along, they can keep watching.

Adam Greenbaum

Yes.

F. Thomson Leighton

And they can. So what do I -- how do I get this?

Adam Greenbaum

Through the same experience. You have to provide some identity so we know who you are and whether or not you have access to the content.

F. Thomson Leighton

All right, so I'm going to sign in now again. All right. Because this iPad is not hooked to the --

Adam Greenbaum

No, there are 2 separate devices. And just make sure you have your identity on both devices.

F. Thomson Leighton

All right, so I'm signing in.

Adam Greenbaum

It's a complicated password for demo purposes only.

F. Thomson Leighton

All right. So I want to watch Buck Bunny. All right. So it's giving me an option here, right?

Adam Greenbaum

Yes. So because we had the same login on both devices, and we're managing that identity in the Akamai cloud, we can keep track of your session state. So we know that you were watching this piece of content and you stopped it around the 1 minute and 15 second mark.

F. Thomson Leighton

Right. So here I stopped at 1:17 and it's picking up at 1:15. So it's a couple of seconds a little back.

Adam Greenbaum

A little bit of rewind time.

F. Thomson Leighton

All right. And so now I press resume, did it get me there? There we go. All right. So now we're picking up right where we left off.

Adam Greenbaum

There we are.

F. Thomson Leighton

Okay, very good. So now the kids can watch in the car. And as we drive around in the car, they can keep watching?

Adam Greenbaum

They can keep watching and go from Wi-Fi to cellular, just as long as the content rights allow.

F. Thomson Leighton

Okay. And now say I get back home again and I want to switch back to the TV?

Adam Greenbaum

Sure. Well, since -- go ahead and pause your content here.

F. Thomson Leighton

All right. Pause. All right, and can you zoom in on that and see where we paused?

Adam Greenbaum

Around the 1 minute and 42 second mark.

F. Thomson Leighton

That's 1:42. All right.

Adam Greenbaum

Go back.

F. Thomson Leighton

I go back to clips.

Adam Greenbaum

Yes, go back to clips. Choose Big Buck Bunny again, and it'll provide you the...

F. Thomson Leighton

And you get to zoom in on that. So we're at 1 --

So we're pick up at exactly the same spot, 1:42, and I can resume the movie from where we were there. So it's going to -- okay, very good. Picking up. Now is this -- this is live today? Customers can use this?

Adam Greenbaum

This is live today and customers can do this.

F. Thomson Leighton

All right. And for the folks out there to see it coming soon?

Adam Greenbaum

Coming soon, we have a major media customer who will be launching early next year. So you'll be able to see this live and in action.

F. Thomson Leighton

So by the time we're all here next year, I would imagine you all -- your kids, at least, will be using this and be supported in the major announcements that are coming early in the year with this are big-name providers that probably a lot of you use for doing this. It'd be pretty cool. All right, well thank you very much. That's cool.

Adam Greenbaum

Sounds good. Thank you.

F. Thomson Leighton

Very good. Okay, thanks. Okay, so that's media. A lot going on there and a major area of investment for us, and one that we hope will provide growth over the next 5 years. I want to switch now and talk about performance. And performance has changed a lot over the last decade as websites changed.

If you go back to Marriott, this a snapshot of the Marriott homepage in 1996. About as static as you get, not particularly rich. And as you watch how it's evolved over the years, it gets richer, more capability, more interactive. You can start doing more things with it. And then you get all the way to today, and if we could put this up on the screen? We switch over -- there we go. So here is Marriott's website today, and it's a lot more interesting to use. Say I want to -- oh, the Caribbean looks nice on a day like today. Let's find out about that. And you can make reservations easily, move around the site, it's interactive, it's immersive. And that's -- let's switch back now, please? That's made the job of the guy who needs that website to go fast be harder. Whoops, wrong button. If you look at the time to download the site, all of the things being equal, as you put that richer content on, that makes the site slower because more bits have to get into your device. And then if you want it to be interactive, you've got all that JavaScript going on and there's inefficiencies there. And that makes it heavier and chattier doing the interactions. That all makes it be slower.

And then sites have evolved to have a lot of third-party content on the site. It's not uncommon today to see a site where half the content is from something else. Maps or ads or who-knows-what, and that slows down the site.

Now at the same time, user expectations have gone the other way. While your site gets slower because it's adding all this content and functionality, users expect it to be a lot faster. Ten years ago, if they came in at 8 seconds, you're doing great. Then the benchmark was 4 seconds. Now it's 2 seconds, going to one second from when you click to where it has to paint. Now you can make money when it's slow, make a lot more money when it's fast. And there's a zillion studies now, a lot that we've done to prove that.

Throw in mobile and things get out of control. And you've all been to mobile sites that take 15 or 30 seconds or more to load the page, and then it doesn't even look very good when you do it. The expectations of the user are that mobile needs to work just as well as a landline. And that creates an even bigger problem today, as there is a proliferation of mobile devices. And of course, it has to look right. It can't be this thing where you're getting a piece of the page on your screen.

So we've invested a lot into devices, in making the devices perform very well. Of course, we've already optimized our network for network communication. We've done protocol optimizations, which we do special things again for mobile. Front-end optimization, which we'll be delivering early next year. And again, special things, in addition, for mobile. Third-party optimization, coming out early next year to optimize all that third-party content on your site, which we might not deliver today but now we're going to on behalf of the site. And also to special things for mobile, where we detect what the device is, and we do special acceleration through DSA to make the mobile experience better. And of course, a lot of folks now are building special-purpose mobile sites where they're using technology to transcode the site to create a mobile site on-the-fly to make it look right.

So here is how the performance has changed using Akamai for the same page over the years. With the various levels of Akamai services over the years, without Akamai, 8 seconds for a typical mostly static page; with Akamai's original CDN service, which has gotten faster over the years, cut that down to a little over 4 seconds; using DSA which is our Dynamic Site Accelerator service, not just content delivery, in fact uses content delivery but that's a small fraction of the benefit, down under 2 seconds; and now with the next-generation performance capabilities that we'll be launching early next year, faster still. And these are the lab tests that we've been doing on that.

Mobile is a different story. Here is one case study we did with a mobile application that included downloading 1 megabyte attachment to an e-mail. And you see that with without Akamai, it took 50 seconds to do; with Akamai, it took 20 seconds to do. And this was a test from last year. And we looked at that test, and we said, "Hey, great. It's a lot faster with Akamai. We knocked off 30 seconds." And that's enormous. You're just sitting there waiting for it to show up on your device. You also look at that and say, "20 seconds is terrible. You don't want to wait 20 seconds. You want that thing right away." So we've done a lot of work, and we can show you the benefits of that work with some of our customer sites today. And to help with that, Pedro Santos, who's responsible for this technology is going to show us how it works.

Pedro Santos

Hey, Tom.

F. Thomson Leighton

What are we going to see?

Pedro Santos

So let's go shopping. We have a camera come up. Let's take a look at what web browsing looks like when Akamai is helping out with the web browsing. And so there's a couple of different things that we want to show. We want to show sort of what Akamai can do to help solve device complexity issues that many companies have today, as well as how we can speed up a mobile version of a website.

F. Thomson Leighton

Okay.

Pedro Santos

So why don't we go to -- it's holiday time of year.. Let's go shopping. Let's go to BestBuy.com. So if you could just type in www.bestbuy.com?

F. Thomson Leighton

Okay. I got it here. I can just get there.

Pedro Santos

All right. Click on it right there. And so what's happening, and you can see how that --

F. Thomson Leighton

It's pretty quick.

Pedro Santos

Pretty quickly. The typical response time you see on a mobile website is anywhere from 10, 15, 20 seconds depending on the specific site. So what Akamai did there is we detected that, that user was coming in from a mobile device, in this particular instance, an iPhone, and redirected that user to the mobile version of the site rather than giving the full www display. Now the mobile version of the site was created by Best Buy themselves, but we're hoping to support them with our detection and redirect technology.

F. Thomson Leighton

And they're using Dynamic Site Accelerator --

Pedro Santos

And they're using the Dynamic Site Accelerator with some special tricks to make it work faster for devices, like being able to cash device-specific content and being able to accelerate the content, whether it's coming over Wi-Fi or over mobile gateways.

F. Thomson Leighton

Okay. So let's buy something.

Pedro Santos

Sure. So I'm shopping actually for my daughter right now as part of the holidays. I'm looking for an iHome clock radio. So why don't you type iHome right in that search box right there.

F. Thomson Leighton

Right up here?

Pedro Santos

Yes. And just hit iHome. And now this -- close.

F. Thomson Leighton

You have a webinar. iHome, did I do that right?

Pedro Santos

Yes, that looks okay. Hit search. Hit the search button. There you go. So what happened there is that is a non-cashable request. Akamai had to go to request that piece of content from a origin server that was sitting in Minneapolis, right? So the Best Buy origin server, and traverse the Internet and then traverse the network that came into this particular building.

F. Thomson Leighton

So that was still pretty fast.

Pedro Santos

Pretty quick, yes. Much better than the 10- to 15-second experience that you typically get over mobile devices.

F. Thomson Leighton

Okay. And by the way, you all can do this yourselves later. Don't all do it now or you'll flood our Wi-Fi connection here or the tower or whatever. So now, let's say which one do you want here?

Pedro Santos

So let's click on that second one right there. She tends to like white, so let's go with the white one. And we'll take a look at that browsing experience.

F. Thomson Leighton

All right. A few seconds, not bad.

Pedro Santos

Two, 3, 4 seconds. And so you get a pretty good sense for what that browsing experience looks like as you're clicking through the different pages. Let's go and -- since we're not going to buy right now, I'd actually like to take a look at it in person. Let's see if we can find a store nearby to take a look at that. So if you click on the store's icon up to the top right, and that particular request was also pretty speedy. Again, we're helping Best Buy with being able to accelerate the request for that particular store. They're doing the mapping to the ZIP code to find the store, and we're helping accelerate that particular request.

F. Thomson Leighton

So it knows where we are and it gave these things in the close-by stores here?

Pedro Santos

It does. Yes, it knows where we are. And you can click on the store nearby, in this particular Dorchester store. And then you click on map and directions, and it'll pull up the Google Maps to go and have you potentially put in directions to get to get to that particular store.

F. Thomson Leighton

That's very cool. All right. So you can now get directions to it on the map. Now I get directions. Wow, so this is pretty fast. So you can actually use the device as fast as a landline and really do some functionality with it.

Pedro Santos

The biggest deterrent right now in terms of mobile websites is that the experience is very, very difficult from a consumer perspective. So as that experience gets better and gets faster and more usable, you expect to see more mobile shopping happening on devices itself.

F. Thomson Leighton

All right. And this is not just a test, right?

Pedro Santos

This is not a test. It's a regular website. Anyone can test it any day. There's hundreds of these types of sites that Akamai helps accelerate today, and there's more work that we're doing to help accelerate these sites both over Wi-Fi networks and over mobile networks specifically.

F. Thomson Leighton

Great. All right, well, thank you very much. That's great.

Pedro Santos

Thanks, Tom.

F. Thomson Leighton

Okay, if we can switch back to the slides? Now in this particular case, I think we're hooked up on the Wi-Fi here, it'll work off the cell tower as well. One of the problems with going over the 3G, 4G networks is that the cellular networks were not designed to do any of this, much less to support videos. They were designed for voice. And with voice, you need about 8 kilobits a second. To do reasonable browsing, you want 1 megabit a second. If you want to do HD video, you're talking 10 megabits a second. So a huge increase in bandwidth consumption. And the networks simply can't support it, and they're not going to be able to change that easily because the connections to the towers and to the infrastructure through the cellular network don't have enough capacity. And those are fixed lines. As we move from 3G to 4G, the technology is changing in the tower, and that'll help give more bandwidth here. But there's bottlenecks that live farther back in the infrastructure which are serious problems. That's why you even have trouble with phone calls in major cities today and keeping them as you move around. Now we're working very hard to help improve that situation, largely by getting our software deeper into the cellular networks. Because if we can be out there, we can decrease the burden on these saturated links, and of course, give a much better performance experience. And we're doing that in part through a major new partnership with Ericsson. Of course, Ericsson manages a lot of these cellular networks and they sell equipment to most all of them, and the idea is to have our software living inside their equipment that lives deep in the cellular networks. And this would allow for end-to-end acceleration and prioritization. And the business model there would be that Akamai would sell an enhanced performance service for cellular networks to our customers that want to pay for that, that have highly valuable traffic. We would then share that money with Ericsson and with the carriers, who could then use that money to build in better infrastructures. And of course, they're also buying the Ericsson equipment with the Akamai software onboard. And this will give now end to end Akamai quality guaranteed performance, and also give prioritization capability for the limited available bandwidth. Because in the cellular networks, there's a big difference in value between the video clip that has lots of bits and is not being monetized, and the transaction to buy the clock radio that has a few bits and a lot of money behind it. The value per bit is probably 1,000 or more to 1 between those 2 examples. And this technology will allow the wireless carriers to monetize that and to get different levels of performance based on the value of the traffic and to prioritize the traffic that really does have value.

We've done some early testing on the technology and it's pretty exciting results. The time to do an application varies dramatically without using prioritization, to the point where you can get de-prioritized and not get service at all if you're really congested. With this new technology, you get a very narrow bandwidth performance and it's very good. And that's very exciting, exciting capability.

All right, next I want to change gears and move into security. Most people, and probably most of you do not think of Akamai as a security company. And this is a challenge for us that we need to change that perception because we have some really great security capabilities. In fact, they're unique, uniquely able to defend websites today against cyber attack. And let me talk about that some.

Let me start by summarizing where we are in internet security and the state of internet security. And this is pretty easy to do. I can do it in one word, I could also do it in 2 words. It's a mess out there. Here are just some of the bullets. This is why I struggle to believe, but the analysts seem to agree that if you take all the money spent preparing for attacks, defending against attacks, all the chief security officer and all those guys, then the cost of the attack and then the follow-up, add it all up globally, its $1 trillion a year now. This is for sure true and even a bigger number right now, that if you're a big commerce site and you go down, it costs you $100,000 a minute. The big sites now, it costs more because there's so much holiday buying. This number is small, and I don't have a 2011 number. And in fact, it's hard to get your arms around this because often we don't know when your credit card was stolen or your bank account was depleted, how your information was stolen. And of course, Internet fraud, they have a hard time accounting for this. But it's in the billions today for sure. It's a big cost to our customers. Because of all the credit card theft, the credit card companies have raised the bar on websites that want to do credit card transactions. And there's something called the Payment Card Industry standard that you have to comply with to be able to use credit cards online. And the cost of getting this compliance for a Level 1 merchant of which there is close to 1,000, is now $2 million or more a year. And then you need to be audited. For Level 2 merchants, it's over $1 million a year. And this is just in the last couple of years that's happened. Big cost there and a big headache.

This is when -- this is probably the scariest of all in some ways. One of the things you really worry about with cyber crime is that the bad guys will get a hold of a major utility, or even worse, a major nuclear utility and do something really bad with it. And it was largely thought that this could never be possible. But of course, we all know now about the Stuxnet virus that was used to compromise the Iranian nuclear facility. And of course here, sitting here, we're all happy that maybe we slowed down the Iranian nuclear effort. On the other hand, I'm not so happy there's a virus running around that's now been demonstrated to be capable of getting into a nuclear facility and doing bad things.

These are relevant examples for our customer base, headlines that you've seen. When Wiki leaks was -- it took down Visa for many hours over a few days. Even worse, Sony lost a lot of credit cards, got a lot of negative press about that. And a lot of companies have been in this position, very bad position to be in. Here's one, if you asked me a year ago could it happen, I would say yes, it could happen, but probably not going to. But it just shows how bad things have gotten over the last year. Rupert Murdoch is having his challenges in the press these days. This is one of his news sites, The Sun, in London. And somebody hacked in from the outside and changed the front page of the website to run this story that says he's dead due to a drug overdose. Some weird drug was found on his kitchen table and his body's found in the garden. And even worse, they ran a non-flattering picture of him there. But here's Rupert Murdoch, his own website, running a story to the whole world, that he's dead. That's about as bad as it gets to have your site corrupted. It's just not what you want to have happen as an IT manager. There's lots more headlines. It's just all over the place.

Yes, it's really ramping up. Paul talked about the stats that in November, we had 2.5 times as many attacks as in all of 2009. And November, we had about triple the number of attacks than we saw in October just this year. Just exploding. This is the total number of attacks measured in 2009. There is 2010, and here's where we are so far in 2011. This is going to go further, of course. The attack sizes are also getting very large. The typical attack size we see is measured in gigabits a second now. And that's just way more than the typical website can afford to provision. They can buy all the best firewall stuff, intrusion prevention systems, go to all the major security vendors and buy everything, and they're still going down because they don't have the bandwidth to support the attack traffic. The big attacks we're seeing measured in hundreds of gigs. A normal company just cannot afford to pay the provision for that. And yes, their carrier will sell them something where if you call them up and say you're being attacked and you're down, they'll work to try to filter out traffic over the next several hours. But it's costing thousands of dollars a minute that they're down. So just the normal way of defending a website no longer works.

Here's the kinds of threats that we're seeing. Denial of Service attacks are just a flood of traffic. Sequel injection, they're trying to get in and compromise the site in some way, steal something or corrupt it. Cross-site scripting, same thing, different mechanism, and lots, lots more, the kinds of attacks that are happening. Basically, everybody's getting attacked and a lot of people are being hurt.

Now there's a lot of reasons why it's exploding. The traditional one was the glory hounds. I'm a hacker, hey, I took down XYZ. But it's more than that now, it's done for profit. You have organized crime operating out of Eastern Europe and Russia. In Russia, you have -- there's several ISPs in Russia believed to be run by organized crime. So forget about trying to get a cooperation there to stop attacks if you can trace it back. There are state-sponsored cyber attacks, where a government is trying to get access to corporate data or to government data. And the emergence of political hacktivism, Anonymous with the Wiki Leaks.

Now we have a suite of security offerings that I want to talk about that stop these attacks. So I'm going to talk about what the new services are, and then give you some case examples.

The first is DDoS Defender, which absorbs the volume attacks. The attacks that are tens of gigs or hundreds of gigs just to absorb that so the customer doesn't get flooded at their data center or their firewall. There's a sophisticated layer of defense, Web App Firewall, which filters out the more nefarious attacks that are designed to sneak in, steal stuff or change stuff to corrupt the site.

There's defense for DNS. A lot of people don't think about DNS when they're thinking about securing their website, but it's the Achilles' heel of any website because it's a small request packet that generates a large response packet. And whenever you have that asymmetry, it's very easy for the attacker. And with DNSSEC, it gets even easier, because now the DNS system has to authenticate the end user and that's CPU-intensive, making, ironically, the job of the attacker easier.

There's a special service which I'll go into in detail that's now live that protects against credit card theft. And finally, we help our customers manage with all the different compliance measures out there. PCI for the commerce industry, HIPAA for the pharma vertical, and then there's special things, of course, for the government sector.

So let me show you how it works at a high level. Without Akamai, you take your application, you stick it in your datacenter or you put it in the cloud, or one of the cloud providers. If you use a cloud provider, typically they don't have firewalls, so you've got to put your own firewall in, and then, of course, you poke a hole in the firewall to let everybody come to your website. That's just how it works. The problem is, the bad guys come in through the same hole you poked because they look like normal people, and they bring you down with a flood of traffic, for example, or they do sequel injection attacks or cross-site scripting attacks that take over your site. Now with Akamai, we provide a shield to the origin infrastructure. With Akamai, if you use the normal routes of going to the site, everybody comes entirely through us. The entire end user experience is local with an Akamai server. Akamai comes back in through the firewall, but now we're authenticated. You don't poke a hole in your firewall anymore for the whole world to come in. You poke a tiny little hole that only Akamai comes in, and you authenticate us to make sure it's not somebody else. So much more secure. The bad guys come, as I speak, we have many customers under attack right now, so we're being attacked. They do take down Akamai servers. They don't get inside them, they don't corrupt them or steal anything. And when they flood them with traffic, which of course they can do, we automatically redirect end users to other Akamai servers so that nobody knows the site's being attacked except our customer when we tell them, and we give them the data that shows they're under attack. And this all happens seamlessly. The attackers can follow the directions to go to new servers, in which case they stop attacking this one, it comes back online and maybe this one goes offline. But given that we have well over 100,000 servers at the edge of the Internet where all the capacity is, we've got a ton of capacity for the attack traffic. And whereas hundreds of gigs is enormous for any of our customers and they couldn't possibly provision that, if they are datacenters and the cloud datacenters, it's tiny for us. When we had our big day of 8 terabits a second, we had 3 or 4 terabits a second of software downloads, so that's more than 10x the largest attack we've ever seen. So the attacks seem small to us even though they're enormous to our customers. And here's where the capacity of our Edge network, the distributive platform, really pays off.

We also offer consulting services before the attack. Most folks do not know how to best configure their site to withstand attack. And if they buy the traditional software and the traditional services, they will go down against the large attacks. It's a certainty. And a new feature is that we will now cover the cost of the attack as part of the insurance with our DDoS Defender solution. We learned this the hard way. We had one of the customers get the 100-gig attack. We kept them up, they were thrilled. This is a high-profile government site, and then they got the bill for all the traffic we delivered, which was hundreds of times normal, and they weren't so thrilled anymore. And so we got the feedback that, hey, you know, it happens really to any customer, just cover it in the cost of the service or charge us a little bit more, which we now do. So we charge a little bit extra and we bear the cost of the attack and they don't get that nasty bill at the end of the year or at the end of the quarter, so very popular.

In addition, we now deploy our Web App Firewall in all the Edge servers; so we actually filter out those nasty requests that are designed to compromise the origin infrastructure way out here before it gets anywhere near the datacenter or even anywhere near the United States in some cases. Much better to get it out, stop the attack before it gets to your datacenter. If it gets at your datacenter in any kind of scale, you're dead.

And we also deploy our DNS service, Enhanced DNS out in those same servers, giving massive capacity for DNS requests and attacks.

Let me show you the special thing we do for our credit cards, which is now live today. Here's the old way. When you're buying something online with the major commerce sites, the way it works today is you're connecting to an Akamai server, generally without knowing it, you give us your credit card. We send it through SSL back to our customer, the e-commerce site, and they store the credit card in their credit card vault. And then the bad guy knows where those are and steals them. And that happens a lot, and is very bad for everybody. Because then those credit cards are used. And maybe your credit card gets used to buy something somewhere else and shipped to some place that is hard to track the bad guy. Now we can stop that. And it uses our Edge Tokenization technology. You still give us the credit card, but now the first thing we do is we send that credit card back to the bank or the gateway that issued it. All right? And they convert it or associate it to a token, which is a random number, truly a random number that looks like a credit card number. And they might include the last 4 digits, which is more open than the whole credit card. And that token now is associated at the bank with this credit card and this merchant. So the token is only good for this credit card at this merchant. Because a normal credit card is good anywhere, which is why people steal it because then they can use it at other places you are not shopping. So banks send us back the token, we send the token to our customer, they store the tokens now in their data vault. The tokens are random numbers. Bad guy comes, he steals them. What did he get? Random numbers, not good anywhere else. They're only good when this merchant authenticates themselves to the bank, which they can do and they can process transactions against this card. So now, the e-commerce site is no longer holding any credit cards. They never touch them. So they don't have to pay all that PCI compliance expense.

So in summary, Akamai protects your site. If these 3 sites, all of whom are customers one way or another, had used Akamai services for security at the time, none of these events would have happened. In fact, at the Sun, they integrated on an emergency basis our security services. And the next day, another attack was deflected and stopped that was going to compromise the site and run another story about Rupert Murdoch, but was blocked because they had Akamai at the time.

Let me go back and give you now cases where customers did have Akamai Security deployed before the attack happened and show you what the result. So we'll start with cyber crime. And cyber extortion is big, especially this time of the year. This is data from last year's holiday buying season. And there was 1 cyber extortion ring that approached a lot of e-commerce sites all at once. They've given the demand letter to send $50,000 or $100,000 to this bank account or we will DoS you. And we had at least a dozen Akamai customers that got the same notice, did not pay and were attacked. And here's 3 of them. And they're unrelated. If I told you the names, you would not see any correlation between them, it's just they were on the target list for this gang for extortion. And you can see how they're related. This is a traffic plot. So here's their normal traffic, which is -- you can't even see because it's so tiny compared to the attack. And then you can see the attack profile is very similar. And they're using a common bot army to do this damage. And you can see the volumes, huge volumes compared to normal, 1,000 to 10,000x normal. No site can afford to pay a 1,000x normal in capacity or 10,000x normal, but they were fine. And as a result, because they were using Akamai, saved a lot of money by not being brought down, and they didn't have to pay the extortion money, which is good.

Here's an example of a state-sponsored cyber crime. This is a case where it was believed to be a foreign government was attempting to steal confidential U.S. government information from a well-known government website. The FBI got wind of it, alerted the site, which did an emergency integration of our web app firewall. And the attack was detected and forwarded. So we can see the entities coming in, block it and the site is just fine.

The glory hounds are still around, and they have new weird ways of getting glory. This is really bizarre. This is the bitcoin attacks. How many of you have heard of bitcoin? Know anything about it? Oh, good, so several of you have. It actually started as an academic university attempt to create a digital currency. Now you think about -- before dollar bills were around, gold was a currency. And gold had value, in part, because its scarce. It's hard to mine gold. Before -- 100 years ago, amethyst was more valuable than diamonds, because it was very little of it until they discovered these giant geodes in Brazil that were full of amethyst. And now amethyst is not so expensive. Well the idea here is to do the same thing in the digital world. And so there are certain problems, math problems, like factoring big numbers, that are hard, computationally hard. And so the idea was that, hey, if you can factor this big number, that has a value. And that creates a bitcoin, okay? Now bitcoins, like a lot of the virtual things today with avatars and electronic gaming, they use sort of a gray market where you can translate bitcoins and sell them for real dollars. The going rate today is about $5. Over the summer, it was $20. So it changes for people who want to own bitcoins. Now it started out innocently, but then -- well, if there's value to it, if you want to mine for gold, people want to mine for bitcoins. They said, "Hey, I'll create a botnet and I'll steal all your computer CPU cycles to do the mining, to factor the big number to create the bitcoin for me." So now it goes into the domain of crime in some sense. And then people took the next step, which is pretty remarkable, rather than go to the effort of taking all your machines and writing codes to solve the bitcoin problem. I'm just going to do cyber extortion. And this is an e-mail one of our customers in Germany got. We've held out the name there, that says, "You're going to get a DDoS attack of 100 gigabits a second unless you donate or give me a 100 bitcoins to my account." Imagine if you got this e-mail, this poor guy had no idea what a bitcoin was or how to get it or how to pay it. So he got attacked, which was fine because he was an Akamai customer. And here you see the traffic graph, classic traffic graph of an attack, and a huge spike in traffic here. Now the weird one about this is the bitcoins then are about $20. So he's only asking for $2,000. Usually, you see a lot more. But the guy didn't have the bitcoins to give them. And then you see the Akamai defense. This is the blocking of the attack traffic because we detect the kind of attack they were doing and it's filtered out. So this is what our customer sees. They see the traffic go up, they see the Akamai blocking it and they see the website performing just fine.

The last one is pretty scary. It's the political hacktivism, which has really merged in the last few years. Probably the most famous example is the Anonymous organization, organization used loosely here, which got its start in attacking the motion picture industry and the record industry because those industries were going after the guys that were stealing movies and songs online through peer-to-peer technology. So the record industry is suing the guys doing peer-to-peer and stealing stuff. And so Anonymous says, "Hey, we represent the people. The people are being screwed by these big companies. We're going to attack those big companies and bring their websites down." And that's how Anonymous got going. And then it expanded to last year with the Wiki Leaks scenario. Now Assange -- everybody, I guess, likes looking on Wiki Leaks to see what the State Department was doing behind the scenes. Our government didn't like it. And they got cooperation from the major credit card companies to not allow people to donate to Assange and to Wiki Leaks to keep the site up. Well, Anonymous thought that was bad because they thought that the Wiki Leaks should stay out there. So they attacked the credit card sites, and they brought down Visa for hours at a time. They've also gone after famous personalities and lawyers and of course security companies were a big target for Anonymous. They've even built their own tool set, which is available online. It's called the Low Orbit Ion Cannon, one of their famous once, or LOIC, has a range of attack capabilities. And it allows you to be a participant in an attack without doing anything. And what you do is you go to the site, download the Cannon, press go, and then one of the Anonymous hackers now can just incorporate your machine into the attack. This takes a little harder to defend yourself when we give your IP address to the government and they come after you. You can't claim that, oh, we were attacked without you knowing it, because you actually had to download the weapon and install it on your machine. Even though you didn't launch the attack, somebody else did it with your machine. Although we got to get people to know us as a security company, maybe you can help with that, because we're not the first thing people think about when they want to make sure their website stays up.

All right, the last area I want to talk about is the cloud enterprise computing and enterprise applications. It's never been an easy story, but back in the old days, it's easier than it was than it is today. You had your corporate data center, and you had a branch office. And you had your app. Your corporate app is sitting in the corporate data center. And you had a private WAN to make sure everything worked well. And maybe you had some boxes to accelerate an offload traffic. Then you started buying companies and adding more branch offices and got more expensive and complicated. And then it really got bad when you have business partners who needed to access those apps online. And now you've got to move in to the Internet. And then you start dealing with end users, maybe for some of those same applications. And then the cloud came along with the cloud providers. And now something happens that's fundamentally different. In many cases, your hardware moved and became virtual software sitting on VMs and cloud providers. So now it's in somebody else's data center, still your app but somebody else's data center.

And the final piece that's happening now is the SaaS world, where it's not your app anymore. It's somebody else's app that you're using and paying for as a service. So you migrated from your data center, still yours, into somebody else's data center. Now it's somebody else's app that you're using as a service in another data center. Now traditional Akamai helps all this because we have our servers all around the Internet. They were making the communications between these payers of entities to be a lot faster and more reliable and more secure. But now we're taking even more steps to get deeper penetration through partnerships. And the first was with IBM, where IBM data power and WebSphere are Akamai-aware and they're integrated with Akamai. And that means now that Akamai lives inside of VM in these devices that will live in the corporate data center. And it will automatically integrate with Akamai services to make the services perform even faster, be easier to use and be more secure.

We have the cloud providers, for example, RackSpace. We move our servers into their infrastructure, into their data centers and integrate with their business processes. So it's easy for their customers to, as they talked about in the customer conference video, rackamize [ph] in Rackspace with Akamai integrated.

And most recently, with Riverbed, doing something really more different and very interesting. Traditionally, Riverbed lives in the WAN. And they accelerate communications between the branch office and the data center, and they offload traffic. They're not useful, though, when you move into the SaaS world because you don't have your steelhead boxes out here. And that means you can't accelerate an offload traffic across the long haul. But we can fix that by working together. And in our new partnership, what happens is, steelhead software lives on VMs inside Akamai servers and Akamai software lives inside VMs on steelhead servers. This now enables our customers to get fast, except get acceleration and improve performance and security from the SaaS provider into the branch office. So you can view this as Akamai being extended into the WAN environment, where we've never really lived, and Riverbed extended into the Akamai Internet environment, where they don't live, bringing the best of capabilities to bear in this hybrid world that's emerging.

Let me show an example, the performance gains you can get. This is a time to download a 9-megabyte file using Office 365. And without the solution, it can take up to 20 seconds. With the solution, maybe 4 seconds. To make a change to the file and upload it, even a more striking difference. Upload is slower on Office 365, but becomes even faster because now we actually have a copy of the file in the library. So it just sends up the change, so very fast. Now it's always nice to be able to show these graphs that we're getting performance improvement. It's even nicer to see it work live and for real. And so to help me with that is Dana Burd, who does all this stuff.

What are you going to show us here?

Dana Burd

All right. Well, today, I'm going to demonstrate the performance and optimization that enterprise might see using this Riverbed Akamai solution that you just talked about. And in this case, the enterprise is using Office 365. Close this down. And essentially, what I have here is a machine in my Cambridge office, that -- it is without the solution enabled. And then I also have another machine with the solution enabled. And then, first, I'm going to demonstrate accessing Office 365 site without the solution.

So I'm going to start up my Internet Explorer here and log in to an actual Microsoft Office 365 account.

F. Thomson Leighton

So this is the real thing?

Dana Burd

This is the real thing and without the solution enabled. So one thing is that the back-end system of Office 365 is not the fastest thing in the world in some days, especially in the morning.

F. Thomson Leighton

And, in fact, is not being accelerated by Akamai.

Dana Burd

Right. So here, we have this file that I'm going to download to the local machine here and transfer that down in here.

F. Thomson Leighton

So we're doing a download now?

Dana Burd

We're doing a download of a file. And as you can see here, we're doing about 5 -- a little over 5 megabits per second download. This ranges during the day, so it could be between 4 and 6. Probably the fastest I see here is about 6 megabits per second.

F. Thomson Leighton

And where is this file coming from?

Dana Burd

This is coming from the Office 365 servers. I just happened to be the instance where they created this SharePoint instance was in Chicago.

F. Thomson Leighton

So that's probably their nearby location to where we are.

Dana Burd

So let me just open this file here and...

F. Thomson Leighton

And how long did that download?

Dana Burd

That took 14 seconds.

F. Thomson Leighton

14. So a little bit faster than I saw, than I put on the slide. Okay.

Dana Burd

Yes. So I'm going to open this slide here.

F. Thomson Leighton

Wait a minute. What does that say?

Dana Burd

So it looks like I have a little typo here, maybe it's the Microsoft auto-correct.

F. Thomson Leighton

Is this my -- this is my presentation from the customer conference?

Dana Burd

This is actually from last year.

F. Thomson Leighton

Last year's customer conference. All right. Yes, let's get that title fixed.

Dana Burd

All right.

F. Thomson Leighton

And now what are you going to do?

Dana Burd

So I'm just going to save this as a new file on my local drive. And then I'm going to transfer that file up, back up to Office 365. So I'm going to add a document here and browse through that file that was created. And I'm going to send it right now. And if you noticed over here, it was about the time that we sent it because there's no dialog box on this particular upload to be able to demonstrate it transferring the file. You can see down here that it's -- the browser is doing something.

F. Thomson Leighton

So it's uploading while we're waiting here.

Dana Burd

So it's uploading. We're about 20 seconds now.

F. Thomson Leighton

It's taking a lot longer to upload.

Dana Burd

Yes. The -- it's still using HTTP upload in this, plus some HTTP posts, so it's a little bit different. Also, the Internet is asymmetric, so your download speeds might be different than your upload speeds.

F. Thomson Leighton

Yes. So it's taking a long time.

Dana Burd

Yes. It's taking even longer than it took [indiscernible]. So typically, about 40 seconds. And here we go, we're done now. It took almost 50 seconds to do that.

F. Thomson Leighton

It's still doing some --

Dana Burd

So this is back-end processing here. The file is actually there. This is the back-end Office 365.

F. Thomson Leighton

All right. So it was 14 down and 45 up?

Dana Burd

Yes.

F. Thomson Leighton

All right. So a minute round trip there for the whole thing.

Dana Burd

Right. So let me just delete this file and do this now with the solution enabled. So now I'm going to switch to another machine that I have with the solution enabled. And here we go. And I'm going to log in to that same account. And now we're still kind of waiting. So it's a little bit faster to load that page.

F. Thomson Leighton

All right. Now we're going to do the download?

Dana Burd

So now we're going to do the same process. We're going to download this file. So waiting for Office 365 to do its back-end processing here. All right. And we're going to save.

F. Thomson Leighton

Wait, so it was already downloaded?

Dana Burd

No. I just saved to this local drive here. And now it's done. So it took just a little over 1 second to download that same file to my local drive.

F. Thomson Leighton

So you went from 14 seconds to 1 second. Now we didn't cache that thing here, right?

Dana Burd

We didn't cache that. So this morning, I actually reset the Riverbed device that was in my Cambridge office. So it did not know anything about this particular flow or this particular file. The way that WAN optimization works is with by doing dictionaries on this metric side. So you have some dictionaries, bits of the bit stream that's going across. But in this case, it was not cached. But there were other performance improvements that we have in the solution, such as -- we have the T3 optimization that the Akamai network also has. We also have the share-out technology, which finds the best path across the Internet. And what's interesting is that even to get to Dallas from Cambridge, going direct --

F. Thomson Leighton

You mean to Chicago?

Dana Burd

Sorry. Excuse me, from Cambridge going to Chicago. Well, the route actually goes to Dallas. And with our share-out technology, we're able to pick that route and go directly from Cambridge.

F. Thomson Leighton

And that gives us a lot higher throughput and means faster for them.

Dana Burd

Exactly. So if we just kind of do the same process here, just to show --

F. Thomson Leighton

All right. So you're making the change and we're going to upload that?

Dana Burd

Yes. So I'm going to just rename it here, same thing. So just to show that this is a new file and upload. Again, these are -- this is back-end processing of Office 365. What's interesting is it just started yesterday. I think they are doing some updates. There's a lot of complaints on the blog.

F. Thomson Leighton

I can imagine. I haven't seen that stuff before.

Dana Burd

So I'm going to click okay now and --

F. Thomson Leighton

Now it's doing the upload.

Dana Burd

Now it's doing the upload. And so it's there. So it's there now, it's processing. The back-end is now processing that document.

F. Thomson Leighton

So it went from 45 seconds to 1 or 2 seconds to upload? That's pretty good.

Dana Burd

So it dramatically speeds up, especially the large files. It does speed up the website itself, as long the back-end is fixed. I think right now, they're just having some issues that they're working on in Office 365.

F. Thomson Leighton

Okay. Is there anything else that we can do with the solution?

Dana Burd

So another example with -- that you might have, another use case that's very common is e-mail, right? So enterprises are now moving their e-mail to Office 365. And one of the things that we showed at the public sector conference was a use case on a field office that might have maybe several users in it. And if an e-mail with an attachment, that same attachment that we just downloaded, that 9-mega attachment, was sent to 5 users in that field office. That e-mail actually goes in duplicate to each one of those users and gets downloaded. So...

F. Thomson Leighton

So Microsoft Office 365 sends it out from Chicago 5x through the Internet.

Dana Burd

Through the Internet. So that exact duplicate e-mail adds up to about 45 megabytes. And on a T1, which is kind of typical for field offices in the public sector, that means you're saturating that T1 for basically 4 minutes or more. And if there's any other traffic on there, on that T1, that gets deprioritized and pushed out. But with the solution, what we do, we're able to do not only speed up the delivery of that, but reduce the number of bits that are going across. So with the solution, we can reduce that down to less than 30 seconds if we've seen the bits before, which is pretty typical.

F. Thomson Leighton

Okay, great. So a company the size of Akamai, if Paul sent out a PowerPoint to our San Mateo office, that would be just tons of -- would just melt them down if they're using Office 365, unless they have the solution.

Dana Burd

Exactly. But with the solution, we're only sending the bits once and then everything else is just references to those bits.

F. Thomson Leighton

Great. That's pretty exciting. All right. Thank you very much. That's great.

Okay. So to summarize, what's Akamai all about? Our core values. Performance has been a key from day 1 and it always will be. We're always making our performance be better. Scale, from day 1, is something we're always working on, trying to get servers in more locations, trying to get our software on every device that's out there. Reliability, we don't talk about it a lot, but we're a company that operates to a 4-9 standard end-to-end downloaded content as measured around the world. It took us years to get beyond 3-9s and an 8. And we've been at 4-9s now for several years. We worked very, very hard at that. Security, we talked a lot about. We've always cared about it, and now we're really making an effort in the marketplace because we have a unique capability to make websites and web apps be secure. Analytics, helping our customers understand what's happening with their site and to manage it better. Support, to give our customers the best care possible to help them manage all the complexities of the Internet. And, of course, the return value on the money they pay to Akamai.

Okay. I think we're going to do Q&A now?

Question-and-Answer Session

Paul L. Sagan

Yes, we're almost exactly on time. Take about 10 minutes of questions on this. I know it's a lot of material. Hopefully, you find it as exciting as we do. We've spent over 10 years building a really unique platform and layer upon layer of technology to bring out new services for the enterprise, for the e-commerce site, for media companies. We hope these demos have given people a great idea of some of what's there and what's coming in the next couple of years. So we're happy to take your questions. Why don't we keep those to technology and products? And if you've got business model questions after J.D. goes, we'll go to second Q&A session. I think there are mics because we're being webcast, if you could wait for those and then we'll get to you.

Unknown Analyst

Great. So you mentioned a lot about the mobile improvements that you're making. I'm kind of curious. Besides partnership with Ericsson, how much of the mobile improvements in technology were actually in-house-developed versus -- I think you did make a small acquisition last year. How much were you able to leverage that for the improvements?

Paul L. Sagan

What you're seeing now is all us. The stuff that was demonstrated when Pedro came up was not Ericsson products and technology at all. We haven't demonstrated that live, although that's now in beta in a couple of markets with different set of products. So what you've seen is a combination of some of what we acquired and a lot of what we've been building in-house.

Unknown Analyst

My question is about the security offerings. I think all of us that watched the presentation were, at least for me, really impressed with what you guys were able to do. And you alluded to the fact that not a lot of people know Akamai as a security provider. So I'm wondering, given the market opportunity, given the offerings that you have, what are you doing to essentially make the potential become reality there?

F. Thomson Leighton

Yes, we're talking about it now. And I think when Bob gets up, he'll talk about that as well, in going to market. We've got great case studies now because we've been out in the market long enough, and people have adopted our services and then proved they work. We have a lot more case studies than I was able to show you here. And I think word gets around now, that we're a place to go to protect your website. And that's our challenge, of course. This year is to really get out of the market place and let people know. And I would hope to see a dramatic increase in adoption of our security products as a result.

Paul L. Sagan

Where is the mic?

Unknown Analyst

I had a question on some of the offerings around getting TV content on PC devices that you demonstrated. And I was just curious through what the monetization model around it. Is it the goal to be more competitive and drive more volume traffic, compete more effectively? Is there a potential for better pricing because of the higher quality of the offering? Or is this something that is considered a value-added service that you can monetize independently of the volume component?

F. Thomson Leighton

I think it's a little bit of all of that. As we do new technologies, we get paid for those new technologies. Also, it helps reinforce Akamai as the go-to company for the delivery and the traffic associated with the delivery and to make us be more competitive in the marketplace. So it's really a combination of all 3. But as we build new services, we get paid for those services, and they're on our platform. So people want to be on our platform even more, and that helps them pay us a better price for traffic. But traffic is traffic. There's a lot of it. And so that's always going to be -- I think you're always going to see pricing come down there going forward as the traffic volume grows. And our goal is to grow the traffic faster than the price drops, and that's how we get growth as a company.

Paul L. Sagan

And one of the other things that's implied but we didn't really call it out in there is what happens through in the network adaptation such as the ability to take 1 file and then have it play either high res or low res depending on the device. It takes a lot of cost out for our customers around transcoding, having multiple formats and then multiple bit rate files and then storing them all. And so one of the things that we believe is we take a lot of the other, if you will, legacy costs of digitizing and delivering content. It allows us in a complete solution to get a better price for what we're doing but just lower our customers' overall cost in the process of doing it. Tim, I think you had one.

Unknown Analyst

Yes, a question on mobile again. Can you give us an idea of where you are as far as the number of deployments with Ericsson you have? And how long you think it will take to get as ubiquitous as you are in the traditional network?

Paul L. Sagan

Well, we're in the first test markets today with the content prioritization service. So that's only in a couple of places and some tests. Now the results are great, and the interest is great. But that's going to be a rollout over the next few years. Some of the services, like you saw demonstrated today, the acceleration is already there because it happens across our network already, because only a piece of it is the wireless last-mile and the rest is the Internet that we already optimized. We haven't made any predictions about the Ericsson specific products, and we won't do that yet. We'll be doing those announcements with them starting next year together. When you say sort of equal to the footprint we have today, that's going to take a long time. In fact, that will never happen because the footprint we have today, we're building that, too. So we're racing ourselves. But I think over the next couple of years, our deployments and technology inside wireless networks will start to look quite comprehensive. You've got the mic?

Unknown Analyst

Yes. So obviously, there's a lot of focus on building out the additional services to help better monetize the network. One of the things that, for example, you mentioned is the tokenization solution. How easy is it for the customers to do these things on their own? In other words, do you actually have to leverage Akamai to create a tokenization-type solution? Or could they actually do that on their own? And then when you kind of think of that more broadly across all the different services that you're thinking about, launching or that you're in the process of launching, what's the likelihood that the customers are going to be willing to pay for those versus just saying it's maybe more cost-effective for us just to create this internal R&D ourselves and then leverage that on a go forward basis.

F. Thomson Leighton

Yes, with tokenization, in theory, they can buy some kind of technology for that or to try to build it themselves, but -- and restrict the number of devices that touch the credit card. But they still are touching the credit card. So they can never by themselves get around that. And if they're touching the credit card, they've got PCI compliance and they got to be audited, they might reduce the infrastructure that's going to get audited. But it's still a problem. So today, if they don't want to touch the credit card and not have that issue, we're the only way they can do that because it comes to us and we take care of that, so they never touch it. So we have a unique capability. Plus now that we've developed it and are doing it for customers, it's a lot more cost-effective to use us than for them to try to reinvent it and then still be touching credit cards when they're done. So I think in that situation, they do need to use our capability if they don't want to be touching it. No, go ahead and I'll...

Unknown Analyst

I was going to say that, just in terms of all the solutions that you're, kind of, in terms of strategically how you're thinking about these services, is -- would you argue that the network is a very critical component to that in terms of how you're going to differentiate the services off on top of it? Or can they simply just create these and you don't necessarily have to have the network for it to be effective for the customer?

F. Thomson Leighton

Well, they need to use us in some form because the credit card has to come to us. And they pay us separately for that service. It's a separate service. It's paid for separately. And it sits on top of the Akamai platform. So they have to be hooked in to the Akamai platform in some way to use tokenization. And typically, the customers will, when they buy one, they'll buy more of the services. So tokenization is something they would add on to the existing service, typically.

Paul L. Sagan

I think the other thing I would add is, these are really hard things to build. And we've been fortunate that we have some of the best engineers in the world doing it. And there's a massive talent [indiscernible] shortage of talents. So one of the things that we hear from customers who aren't experts in this, right, they run commerce businesses or banks or airlines, is they have very few developers. And a very limited capability to do anything from an engineering point of view. So they are really looking for a cloud provider to bring them best-of-breed and make the integration work the only thing they have to do and even make that as simple as possible because they can't build these kinds of solutions in-house. Even if they could, I don't think it would be economical for them to do it. But they simply can't even find the talent to do it. You've got the mic?

Unknown Analyst

Yes, on the video side, are you doing the transcoding on the edge? You aren't. So what does that imply for certain servers and storage and all that sort of stuff as you have all these different versions? Are you doing it dynamically or beforehand?

F. Thomson Leighton

We can do both. It implies less storage as needed. Because if you multiply the number of different formats by the different bit rates, you can get a big number. And that's one of the things that makes it expensive for our customers. They've got to store it in 50 or 100 different ways. That's expensive for movies. And we can do that on the fly. Saving storage, it uses CPUs that may be otherwise idle to do it. And it's done throughout our network today. Yes, Michael?

Unknown Analyst

On TV Everywhere, I'm trying to get a sense of how big this opportunity really is. Are the NSS primarily interested in this for out-of-network delivery? Or are they interested in for in-network delivery and in-home delivery as well? And if so, why do you add value there over what they can do on their internal networks?

F. Thomson Leighton

Bob, we make it be seamless, and that's what adds the value. So you can go from the in-home experience, as we demonstrated on the TV screen, to the car driving around somewhere else. And you can pick up where you left off and all the stakes maintained and all the rights management is maintained in our cloud. And it can move from 1 provider to another. And that's where this provides value. We also provide separate value within a given provider's network, accelerating it and reducing costs and making performance be better. But TV Everywhere is about moving from 1 location to other locations with other providers in other devices and context.

Unknown Analyst

But they're interested in using you for in-network, as well as out-of-network delivery?

F. Thomson Leighton

Yes.

Unknown Analyst

Notwithstanding the tremendous talent pool within Akamai to work on things like security, have you given any thought to working with any of the more established security vendors to promote your products in this area? And I guess, more importantly, are any of your large customers asking you to work maybe more hand-in-hand with some of the more established security players?

Paul L. Sagan

That's come up a lot, and we have a lot of those discussions. But their models are really different. They're either defended inside the data center or on the desktop. And those models are changing pretty rapidly to cloud-based models. So we've looked at collaboration. We're certainly compatible with a lot of what goes on. We've spent a lot of time, for example, with the AV companies over the years to make sure that they didn't, if you will, get false positives because they were reading Akamai not the domain they expected around certs and things like that, so that we have seamless operation. But today, we don't have joint offers. We certainly look for them and welcome them, but we haven't found that as a successful path even if it seemed likely.

F. Thomson Leighton

Yes. As those guys move to the cloud, increases the chance of a partnership there because we're a natural platform in the cloud for them.

Paul L. Sagan

And you can think about filtering all kinds of traffic at the edge and doing things with that intelligence based on what a security company might know and be able to apply somewhere else.

Unknown Analyst

Very impressive, the innovation and the pace of innovation, how you're keeping track. Say a little bit about your sales force in terms of how they're distributed across the client base, how you're keeping the customers updated on what you have to offer and the whole upselling effort as you have these innovations and these stories to tell that are very compelling, at least...

Paul L. Sagan

I think that's a better question for when Bob comes up. He owns that around the world. And if he doesn't touch on it, we'll come back to that, for sure.

Unknown Analyst

One of the products you didn't talk too much about, I think you guys are working with the UltraViolet coalition on UltraViolet DVDs. And on the Harry Potter release in the fourth quarter, it was so problematic that Warner Bros. ended up giving free iTunes coupons to the customers to download that. Can you talk about what's going on with that product.

Paul L. Sagan

Well, we won't talk specifically about that example in the market place. But that's another form of the idea of TV Everywhere. Because the idea, if you have rights to the content, can you get it on any device? And we support UltraViolet and the same kind of technology we've applied to UltraViolet solutions. Somebody did a big rollout this week, and we were a part of that and effectively manage the authentication and delivery between you, the locker, your rights in getting the content back to it. It's another example of trying to say, "Can we have 1 repository of this content?" And then if somebody has the rights, you ought to follow them regardless of what network or what device they are on. And then they ought to get the appropriate experience to the big screen or the small screen depending on the connectivity they have in real-time.

Unknown Analyst

You've shown a lot of products that all seem to be best-of-breed services that if you look competitively, you may be competing against a lot of very small segmented products. To what extent, as you look at this, the content delivery assurance, payment security, are you pursuing a vision of being a full service provider offering, say, a turnkey solution where you are able to really lock in customers and protect yourself competitively from competition?

Paul L. Sagan

But I think that's a good preview of some of what Bob will talk about. But very much, we think about people becoming an Akamai shop. And that's the kind of vocabulary customers have started to use in places like the conference because we're really bringing them a platform that they can continue to build things. So the first customers who are delivering a few videos 10 years ago didn't think about security authentication, TV Everywhere, adapted bit rate streaming, multiple devices, et cetera. But they expect us and they push us very hard to have a complete solution that's been extensible to new things. And you're right in many of these markets. We do compete with small-point providers. We try to say, "I'm almost as good, maybe I'm even better at one thing." Increasingly our customers say, "That is just too complicated. We cannot deal with all these solutions and integrate them into a live dynamic environment that's moving so rapidly." So we think one of our advantages is the ability to wrap things onto a seamless platform, often stuff that we build or supporting a standard like UltraViolet or sometimes integrating some other solutions into our products as we've made some announcements, for example, in some content optimization where we've taken some third-party capability and embedded it into our network as well.

Unknown Analyst

Just as a follow-up, are there partnerships or acquisitions or pieces of the puzzle that you might need to fill in, in order to be able to offer a --

Paul L. Sagan

We've done -- Yes. I'm sure there will be and there have been in the past. So we started in peer-to-peer, for example, by acquiring a very small technology company. Now with the technology that Tom demonstrated running in Russia is pretty different than what we bought, but that started. And some of that kicked us off. And moving in that direction, we'll continue to look for those kinds of things. And in some cases, we've done acquisitions that were more market consolidation. And in some cases though, for example, some of the acceleration, the protocol Tom talked about, some of that came from an acquisition. So we're always looking at both. We'll take 1 more question, and then we'll make sure we get a coffee break or a bio break.

Unknown Analyst

You briefly mentioned the upcoming launch of next-generation performance solutions. And could you elaborate on how significant are the improvements and the performance and what the timing and which products they were talking about?

F. Thomson Leighton

So this would be early next year, be the DSA and raw [ph] product lines. And I'm optimistic for a significant improvement in performance, improving our existing capabilities in communication there and adding new capabilities on front-end optimization, third-party content optimization and a special focus on mobile because you can get extra bang there.

Paul L. Sagan

Great. Thank you. Why don't we take 10 minutes? Try to be back in your seats at 5 after. There's time for restroom breaks, a little coffee. And we're all going to be here if you want to continue these discussions.

[Break]

Operator

Good morning, ladies and gentlemen. Please join me in welcoming Senior Vice President of the Akamai Platform Division, Robert Blumofe.

Robert Blumofe

Hi, everyone. Okay. Why don't I jump right in and get started while you all are getting your seats. Getting right into it. You already saw from Tom some of the range of Akamai solutions that we bring to our customers to power their e-businesses. And what I'm going to speak about is what we call the Akamai Platform or the Akamai Intelligent Platform. And this is a -- essentially a layer of software, of intelligent software that lives at the core and powers all of Akamai's customer-facing solutions. You can think of it as providing core functionality such as mapping the Internet. The platform actually builds real-time maps of the Internet, constantly being updated, identifying, again, in real-time, which paths on the Internet are working well, which paths maybe are having problems, maybe experiencing congestion, maybe are misconfigured, even. It is constantly looking at which servers in our network are heavily loaded, which ones are lightly loaded, which ones might even be experiencing problems, whether it's a hardware problem or a power problem or something like that. It's also providing capabilities such as managing the configuration of a network, installing software across our entire network of servers, managing the configuration of that software, monitoring the software in real-time, what's going on, what's working well, what isn't working well. Let's see. What else is it doing? It's collecting data from all of those servers to provide views for our customers and to everything that's going on. And the list goes on and on. It's a large collection of core capabilities that all of our products depend on. It is -- I was about to say the combination, but it's not the combination because we keep on going. It represents decades of research in distributed algorithms, really research that goes back even before the founding of the company, research that personally I was privileged to be a part of in my days as a graduate student working with Tom. It also represents countless person-years talented, as Paul mentioned, talented person-years of software development, implementing algorithms, iterating on algorithms, throwing out the old and implementing the new, constant innovation, improving the capabilities of the system.

It is an asset that is clearly unique. There's really nothing else like it in the world, not to mention, in the industry. And I would say it's unique in a number of dimensions. It's certainly unique in terms of its reliability. It's unique in its performance. It's unique in its security. And it's unique in its efficiency, its cost effectiveness. And it's certainly unique in its unprecedented scale, some of which Tom already talked about. And one way to illustrate the scale of the platform and some of the functions that it performs is to look at what happens every 60 seconds. So in 60 seconds, the platform will execute some nearly 9 million lines of code. It will process 2 billion loglines, representing transactions that our customers' customers are executing online. It will perform almost 5 million active SureRoute traces of diverse pads across the Internet. It will serve 45 million DNS requests, and it will perform the 60 million server assignments and on and on and on. It's operating at unprecedented scale. And it's really designed for this scale and it's designed to go even well beyond this. Tom showed you some of the numbers. He showed you that really we should think of even these numbers as potentially very, very small numbers. I've been at Akamai long enough to see these as huge numbers. Because compared to what they were just a few years ago, they're really extraordinary, frankly. But I'm also well aware that the potential is such that -- and the expectation, frankly, is that these numbers are going to look quaint in a few years. And one of the unique things that we have in the platform is the ability to keep on going with these numbers, to have these numbers double and double and double and keep going like that and still be able to handle all that load, do it reliably, efficiently and so on. All those things.

So as a software layer, it's truly a unique asset. It runs on distributed platform that is also a unique asset in the world which is what we call the Akamai distributed network, and this is the physical network, the actual servers that are deployed all around the world. It's a bit over 100,000 servers today. And actually, I didn't even put that number on the slide, mainly because to me, that's not what's really important. In fact, I think Tom might have alluded to this issue that we actually worked very hard to keep the number of servers down, right? We want to be able to manage the workload and do everything that we do with the fewest number of servers possible. And we work very hard to do that. And I think it's safe to say that we're certainly, from a server count point of view, not the largest network out there, but it's also safe to say that we're the most distributed computing platform out there. These numbers are, I think, speak for themselves. These servers are now deployed in 1,046 networks, 1,910 locations, 75 countries, ubiquitous, all around the world. These numbers today are at an all-time high. You can see the growth just over the past year, but these numbers are at an all-time high. Our network relationships are at an all-time high. Not just in terms of count as you can see here, but also in terms of quality.

I've been at the company pretty much from the beginning and it's safe to say our network relationships are at an all-time high in terms of both quality and quantity.

And the reason we're able to do this, that is have all these network relationships and extend into all these different networks is because the way we do things, the fact that we deploy our servers in networks, as Tom talked about, and rather than just connecting to them from the outside. The fact that we go in the networks, the fact that we're distributed means that our technology platform and our business model is inherently synergistic with the networks. And what I mean by that is that we actually help them address some of the biggest problems that they are all facing. And to be clear, the networks really are facing significant challenges. In fact, I use this picture quite often, because if you walk into a bar and you sit down next to a telco executive and you say, "How's business?" chances are they're going to reach for the closest cocktail napkin and draw you this picture.

And what this picture shows you is the fundamental problem that they're all wrestling with, which is that revenue per subscriber is roughly flat, but the cost per subscriber skyrockets. It continues to go up and up and up, and that is because the subscribers are doing more and more online. They're shopping more online, they're interacting over social networks more online. They are watching video more online. Everything. And that means the traffic goes up, up, up, the demands go up, up, up, and that means the cost goes up, up, up.

And one way to see sort of what underlies the cost structure is to take a very simplistic view of how a network operates and what it does. And at its core, you can see in this picture that an ISP fundamentally serves the purpose of connecting subscribers, that is their customers, it's you and I at home, to the Internet. So at the bottom of the picture, you can see the subscribers. And at the top of the picture, you see the Internet. And I've labeled this slide as the Oversubscription Bottleneck. And the reason is because there's a lot more capacity at the bottom than there is at the top. Another way to say it -- and I think Tom actually made the statement, is that there's a lot more capacity at the edge of the Internet than there is in the middle. And when I say a lot more, I don't mean like a factor of 2, I mean multiple orders of magnitude difference. There is way more capacity at the bottom of this picture where the subscribers are than there is at the top. That's the oversubscription. And it kind of works where the networks can kind of get away with this for the same reason that oversubscription works in other businesses which is that not everybody shows up at the same time. Not everybody, all the subscribers, use the Internet at the same time. Some of us are doing other things. Some of us are using it, some of us are not. And even if you and I are using the Internet at the same time, at least the old Internet had kind of a burst-and-read property, right? You clicked on a link, there is a burst of activity and then the activity would stop and you'd read for a little while. Then you click on a link and there's a burst-and-read. And my burst and your burst might happen at different times.

But the new Internet, the new web doesn't behave that way. The protocols are chatty, we're constantly interacting. There's constant refresh and change going on. And of course, in the case of video, there is a constant stream of very high bandwidth bit flow going on. It's just a continuous flow. It's not burst-and-read, to say the least.

The other thing that goes on, of course, is that for some reason we live these remarkably synchronized lives, we all tend to do things at the same time, that's why we have traffic jams. That's why you can't get a reservation at a restaurant on Valentine's Day. Why we all celebrate Valentine's Day on February 14? I have no idea. But the fact is we live these remarkably synchronized lives which means that people watch video at roughly the same time. So this oversubscription benefit that the networks get is kind of evaporating. Well, it is evaporating. It's evaporating very rapidly. And so for an ISP to address that growing traffic that we saw in the previous slide, that picture on the previous slide, means that potentially they have to build out extraordinary amounts of capacity within their network and connecting them to the Internet. And that is essentially cost-prohibitive.

But what we can do for them is by putting our servers in their network. And not just in their networks staying in 1 or 2 locations, but in their network in all of the major metros where they have subscribers, means that the subscribers are getting their content from servers that are in the ISP, generally in the same city as that subscriber. So all of that bit flow, all of that video, all of that chatty traffic no longer has to traverse the links that connect them up to the Internet and no longer has to traverse their nationwide backbone. It's delivered locally. That takes load off of their backbone load -- load off of their links up to the Internet, and that allows them to scale far more cost-effectively and address that growing traffic far more cost-effectively. It also, of course, has the added benefit that their subscribers get a much higher performing experience on the web. Because as those subscribers are accessing websites that are Akamai customer websites and that of course, represents much of the most popular content and most popular applications that the subscribers are using, they get a higher performing experience as you saw in some of those demos.

So there's tremendous benefit for an ISP having servers in their network. And that's why we're able to cost-effectively have so many network relationships. And this is something, by the way, that we've been doing for a long, long time. Really since the beginning of the company, we have used this approach, and we have provided solutions to the network operators based on this idea of having managed servers in the network.

And what I'm showing you on this slide at the top is the 2 managed services that we provide today and they are what we call the Akamai Accelerated Network Partner program and a managed CDN, an MCDN program. and they're both based on Akamai managed servers being deployed in their network. And the distinction is simply that in the case of AANP, the servers are there for the purpose of delivering Akamai customer content. And in the case of MCDN, the servers are dedicated to the network for the purpose of delivering their content, which could be content that they have licensed from partners and that they're selling in packages to their subscribers. It could be content that comes from their customers. Or whatever it'd be, it's dedicated to their content. But it's roughly the same idea of managed servers in their network.

From the bottom half of this slide, I'm showing you new solutions that we are in the process of developing and will be bringing to market soon based on licensed software. And I'll tell you a little bit more about those solutions in a minute. But before I get to that, I want to just make one more comment about the fact that all of these solutions are based on highly distributed infrastructure. They are based on the same technology platform that enables highly distributed infrastructure. And to be relevant to the telco, to the network, this is paramount. This is absolutely key to the relevance -- to being relevant to the network. And the reason for that -- or understand that it's worth taking just a little bit of a more detailed look at what these networks look like, how they're architected.

And the purpose of this slide isn't to explain the alphabet soup of infrastructure, but simply to make the point that a ISP, in this case a DSL ISP, is comprised of an incredible amount of infrastructure that is incredibly distributed. They have expensive infrastructure deployed very near their subscribers in a highly distributed fashion. And much of their costs is at the lower part of that picture. When you look at this picture, the big part of their cost is at the bottom part of the picture near where the subscribers are where that infrastructure is most distributed. And mobile networks are the same idea. The infrastructure is somewhat different. It's a different set of acronyms, different alphabet, but it's the same idea, highly distributed infrastructure with a lot of cost very near the subscribers. So if you really want to be relevant to the networks, you have to provide a highly distributed solution. They demand a highly distributed solution. So when we work with the networks and deploy servers in their network with our managed servers, for example, we are always looking at how deep into the network. How down, how south in this picture can we get. And we work with them to deploy our servers in as distributed a fashion as is possible. And in many cases, we are able to deploy in many, many locations across any given ISP. But it does occasionally arise that for some ISPs to go as deep as they want us to be, and they want to be very deep in the network, to go to that level, they require a licensed model, a licensed software model.

And that's where the offerings that I'm about to introduce come into play. They're all based on licensed software, and they're based on a new component that we call the Smart Connected Cache. And the Akamai Smart Connected Cache is a software component. It combines both transparent caching, as well as an Akamai Edge server. And it's sold at a licensed model. It can be embedded in a third-party hardware devices. And it provides greater offload for the network than they can get with any other solution. And that's for a few reasons. One reason is it can go deeper in the network than any other solution. Another reason is it can effectively cache Akamai content, content that otherwise would not be cacheable if they use any other solution. And the third reason is that Smart Connected Cache is smart and it's connected, meaning that it learns from the global Akamai network about what's popular, what's not and makes more intelligent caching decisions, therefore gets better hit rate, therefore better offload. So it can provide greater offload that any other caching solution out there.

The LCDN or licensed CDN solution is a complete CDN solution based on the Smart Connected Cache and provides monetization opportunities for the network, that is they can use the LCDN solution like they use the MCDN, the managed solution today, but now in a licensed model to monetize content.

And finally, all of these solutions, whether it's the managed CDN or the licensed CDN support what we call Global Federation. And this is an Akamai solution to the federation problem. Some of you may have heard this term being used. All of the networks, as they look at content solutions, all require a solution that is not only for their network, but also those off network. And we had this question earlier. There's no such thing as a relevant solution that is just on net. Solutions require off-net capabilities as well. And there's a lot of conversations going on around this idea of federation. Outside of what we're offering though, it's all at the research stage. And I think it's safe to say that best case, those other federation solutions are looking at years before they are working. And even when they are working, they're probably just supporting the most basic delivery functionality, think of it as maybe Akamai circa 1999 or something like that. We have a Global Federation solution that works today. It connects their CDN to the Akamai global network, provide a global solution that they can't get anywhere else. And also supports advanced capabilities that are not available anywhere else.

So to conclude, I just want to leave you with these 3 key points. First, the operators are absolutely clear on this point, they require highly distributed infrastructure. Not just somewhat distributed, highly distributed infrastructure, to the extent that there was ever a debate about distributed versus centralized. The answer is in, the networks are clear. The only way to architect for the Internet and be relevant to the ISPs is to support highly distributed infrastructure, a highly distributed platform. We've been doing this, working with the telcos for a long, long time since the beginning of the company. We've been doing it with managed services, and we're in the process now of expanding the portfolio to include license capabilities.

And finally with these solutions, we are enabling our network operator partners to be more active participants in the content ecosystem. And that's really a good thing for everything because that's going to raise the whole ecosystem and allow for more use of the Internet, and that's just good for everybody. So that is -- those are the key points.

That is the end of the platform presentation, and I'll now turn this thing over to Bob Hughes. So let me introduce Bob. He is our Executive Vice President of Global Sales, Services and Marketing. So with that, Bob, you are on.

Robert W. Hughes

Thanks, Bobby. It's clear that the relations with our networks have never been stronger. And it's absolutely critical to our go-to-market strategy that Paul outlined about how do we achieve the $5 billion milestone. It's -- distribution is one. We have to be more distributed, and Bobby talked about those strategies. What I'm going to do today is I'm going to take you to do kind of a customer’s view. I want to share with you the key market conditions and trends that are driving our business. I want to highlight a few examples of how customers are adopting the platform and leveraging Akamai as a complete solution. A few of you had asked this question, and it does have a significant impact on the go-to-market. And lastly, I'm going to go into more detail in our go-to-market, because there's some significant changes. One of the biggest ones is being this solution partner ecosystem that Paul talked about.

So as Paul talked about earlier, the hyper-connected world is driving the market and changing business dynamics. And we believe there are 4 key market conditions that are impacting this transformation. Companies are turning to the cloud to reduce cost and increase productivity. And now more than ever, these apps have to perform. You heard Tom talked about it. These apps are also being accessed by an increasingly mobile workforce and its global. The workforce is becoming more distributed. The world of media entertainment is radically changing. The appetite for HD video continues to increase, and consumers expect a high quality experience on any device anywhere. In mobile, we're seeing an absolute explosion in a number of connected devices. Today, the ratio of devices to Internet users is 3:1. Analysts are now pointing to a ratio of 5:1 by 2015. Our studies are showing that we believe it's going to be 10:1 by 2020. So again, think about the problem that's going to start to cook. All these devices in the homes that you saw on the network -- on the charts. And then as the users have become more global and more global.

And so as a result of so much business and entertainment moving online and Internet becoming so integrated into every single aspect of our lives, the security threats have become much more frequent, much more dangerous. No longer is security just an IT issue. It's become a serious business issue. And as Tom demonstrated, the attacks we're seeing can literally cripple a brand and cost the company millions of dollars. And as a result of these dramatic shifts, we're seeing an increase in the adoption of Akamai services.

So let's talk about this for a minute. So as it relates to the cloud, we now have almost 600 customers leveraging our Application Performance Solution which is a 50% increase in just the last 2 years. Likewise, we've seen a 66% increase in the adoption of our Dynamic Site Accelerator service which is now used by over 1,100 enterprises. In the media space, we have over 500 customers now using the HD network. Again this is rapid adoption of a service we launched less than 2 years ago. Likewise in the adoption of mobile, our mobile solution portfolio has grown -- there's now 3,000 Akamai's mobile sites up from just a few hundred last year.

On the security front, Tom talked a lot about this, we've seen a tenfold increase in the number of DDoS attacks since 2009. And today, we have over 300 customers buying our solutions. And as you mentioned, the world does not yet know we play here. We've had to protect Akamai's customers. The first 13 years, if you will, of our existence, we built security into the platform from day one with the vision of take those trends, look at the platform. We started to think what are the customers going to do online? What are the enterprise are going to do? We have to protect it. So it's only now that we're actually really starting to take these products to market. And it's something that customers are frankly just asking for. We're turning it on. We've over 300 using it. And really, less than 2 years of us even announcing it as a product, and we haven't gone wide yet in the market.

So why are customers turning to our platform? Because our platform, at the end of the day, it provides a complete solution that addresses all the challenges businesses will face online. In the past, we talked about the industry go-to-market. It's all about understanding the customers, understand what they're trying to do, your roadmap is going to become very clear. As the workforce -- as everything moves online, the type of questions asked is will this app be safe, will my enterprise be safe, will the user be safe? Will it work on multiple devices? Because I'm not -- all the different connected devices that are out there, think of a format, think of what's happening, think how many you have. Is it going to work? We will perform globally at scale because it's going to go out on the Internet and instantly we'll become a global player. And I don't want to have an infrastructure to do this.

So what differentiates Akamai from a traditional CDN or how we've evolved from this traditional CDN. This is a true cloud platform. Too many times people think of the Akamai network. Hopefully, one of the things you took away from this morning's talk is the network. Then there's this software platform that rides across. That's where all these solutions are being built upon this true cloud platform. Our platform is the only one in the world that really enables enterprises to deliver a secure, high-performing user experience any device anywhere in the world. Which brings me now to really, let's get into the weeds a little bit here and talk more about the solutions. So you see plat -- so you see the trends. Now you got those -- the platform. On top of this platform, and this is a big change for Akamai, we're building on top of the stack. Okay? So we now go-to-market with our application and cloud performance solutions, which empower businesses at the end of -- manage cost and securely deliver my application. This is the APS product line, this is our cloud OEM products. Our Website Optimization Solutions enable fast, personalized and secure sites at any device, anywhere. So think about the enterprise. They're now talking the back end apps, and the front and the website. That's where you've traditionally seen us play. Our video and software solutions allow businesses to engage audience with an HD quality video experience and deliver superior software download experience. So again, think of the enterprise and all the different things that are starting to occur on these sites. And as Bobby mentioned, our network -- our recently announced Network Operator Solutions help us expand our network, enable the operators to reduce costs, instantly scale and generate revenue, which benefits all of these products and the platform, and ultimately the end user.

So this whole portfolio is designed to help customers address all of the challenges that come with moving more of their business online. And today, regardless of what industry we've talked about in the past, the customers are starting to buy all these solutions to meet the demands of this hyper-connected world.

So let me give you a few examples. So in the past, we've talked about a media company, the News Corp. We would, of course, go in. The first thing we would sell them was the brand site to get the site up and running. Of course, they're a natural player for video. But now as News Corp. moves it content behind a paid firewall, is it a commerce site, or is it a media site? And now it needs all the commerce solutions. It needs a DSA. It needs to be compliant. It's now going to be Tokenization. It's going to be in TV Everywhere. It's now basically using all of our products -- they're using all the products across the solution line. And so a big part of the strategy going forward, is of course, going to continue to add new services but watch what the customers are doing. There's a huge untapped opportunity for all of these services.

So as all the media companies will face the same challenges, they move content behind the paid firewall, high-quality video on a commerce site. They need a lot of analytics and they better protect it. Okay? So now it's a whole, whole different experience for them. If you think about point providers in the market, there's no one else in the world that could walk in and have that conversation. We might face a competitive threat or someone says, "I could speed up this piece of your app." There's going to be a conversation that these serious enterprises, who we go after? The Fortune 1000 who they're -- what they're thinking about? Well I'm going to point solutions that's going to do it. I'm going to put my content online, any device, anywhere, securely. All has to happen seamlessly. And through that platform, I need to understand it. I've got to see the help of the Internet, and I'm going to see the help of all these applications performing.

Another example is online banking. One of the biggest challenges facing the banks today is this huge transfer of wealth. If I go to my grandparents, they knew who the bank was. They knew the building. In fact, they walk into the bank. One of the concerns that's on every CIO and CTO we meet in the bank are now saying, "Our biggest fear is an iPad" because the bank is now defined by its mobile experience. So in the past, when we said we sold in the financial services, we sold naturally our APS solution. Take a bank like Standard Chartered, they started there. They then deployed our DSA solution. Now the concern about the bank better appear well on a mobile device. So now I need mobile adaptation, mobile redirect. I'm going to expand in new markets for making my bank global. Now as Tom mentioned, this whole thing better be secure. So what company can walk in and pitch that? Only Akamai have the services team to back that up.

Let me give you a few more examples of how customers are transforming and reinventing their business online with the help of the Akamai platform. Let's take Harley-Davidson. Harley is a great example of an iconic manufacturing company that reinvented its go-to-market strategy and transformed its business by opening new revenue streams online. They've now launched an online-only clothing brand, thanks to the Akamai platform. So they're going after the non-rider community, trying to create more market awareness for the brand. They've increased motorcycle sales and brand loyalty because now you can go online and configure your own Harley from over 4,000 unique parts. Okay? So think for a minute about all the applications that are going to drive that behavior. The site's going to perform, I'm going to build it, it's going to interact with different suppliers. I'm going to order it. Then they went -- so we've got to ramp up our other brand even further, so we start to use social media connect with these next generation of riders by partnering with gaming [ph] companies to extend their brand online and rolling out crowd sourcing initiatives on Facebook to build the next generation of advertising campaigns. So again you take Harley, throughout this economic downturn, their sales are up 5% which they directly attribute to what they've done online. For Akamai, going to the manufacturers was not the traditional market we went hard after. Sure, we went into automotive, but if you think about the supply chain and think about what they're doing now, suddenly is this a commerce site? It is apps? It's got video, it's got social media. Every product we sell is suddenly important. We haven't even talked about what's going to happen with the internal apps that we just saw at Riverbed demo. Okay? So again representing this is a huge opportunity for Akamai, these businesses completely transform themselves online. And in the consumer Internet space, we're witnessing a tremendous transformation going on. Invention of complete new business models, customers of ours like Facebook, Groupon, Guild Group and others have demonstrated that there are revenue opportunities for our services -- for new services that are being created. The fact that we, as consumers, didn't even know we want it yet.

And relatively new to the scene and good example, this Airbnb. This is a community marketplace where they enable you to rent out your home to travelers looking for an alternative accommodation. In just 3 years, they've reinvented the way we could think about hospitality. They have properties now in 2,000 cities. They have over 100,000 listings worldwide. They now represent more rooms online than most major hotel chains. And they did this instantly by joining and partnering with Akamai. Think about this. You come onto the scene, they turn on our services, instantly, they're a global player. What's happening on that site now? You want to see pictures of the room, you might want to see a video, you can interact, you're going to book, you're going to do commerce. So again, is it a commerce site, is it a video site? If you think of the target market, what are they looking for? The complete solution and the complete solution done securely.

We talked about video and the growing demand for online video is fueling revenue for sports programming. The NFL is a great example of a media company leveraging the Akamai platform to meet the demands of high-volume traffic and to capture new opportunities, additional users to supplement what they're doing on traditional broadcast. They're trying to engage new audiences online.

What I'd like to do now is I'd like to invite up Noah Fischbach to share how the NFL is growing its fan base by providing a unique, engaging viewing experience online. Noah?

Noah Fischbach

Thanks, Bob. So I have a fortune of joining the league in 2006. Right as we're making a strategic shift in our business. Up until that point, we had third parties run our site and run sort of our digital assets. In 2006, we decided we want to bring all that stuff in-house. And a key portion of that was video. We said, "We need to do something with video." We in the previous site, previous incarnations, we haven't been doing anything. We haven't done -- sort of undeserving that market, if you will. I think part of that is because the technology probably wasn't quite there at that time. And so in 2006, we said we wanted to make video sort of a key differentiator for what we're going to do. And it was -- it's how we were going to differentiate ourselves from the competitors in our space. And so we said, "How can we put video on every single page of the site?" And we went ahead and did that with partnering with Akamai and it's been fantastic since that point. And what was interesting is that as soon as we started putting all these video content out there, our fans continued to ask for more and more of that content.

So today, we don't just put one clip of the game. We put hundreds of clips of the game out there. We slice it up into a clip for a particular play into a clip for a particular player into a clip for a particular moment. A cheerleader being eaten by a mascot, put that out there. And the consumption of all that content has been tremendous. So we've really thrown out -- we've really tried to get a lot of our content out there and bite sized it up in any way shape we can do. And what we've seen it's been incredibly additive to the marketplace where it's continuing to grow our consumption across the board. TV ratings are up and our consumption is up.

And in addition to the amount of content opening out there, our fans have been asking for high fidelity, high quality of content. I think it's no surprise that they're asking for that. Most fans are at home working on HD TVs. And they're expecting HD-quality content. They want that big content and make it look great. And it's kind of something we need anyways because our game moves so fast. You need to be able to see the clip in high fidelity to really understand what's going on. And we continue to do that, working with Akamai and today. And so we don't just put out one clip of a particular piece of content. It is multibit rate. It is multibit rate from a very low format, depending upon a particular device to a much higher bit rate, depending upon if you're in front of your desktop where you can get that full web stream megabit, stream coming onto you. And it's great.

One of the other things that we've also been working on with the help of the Akamai is timeliness of content. In 2009, we began -- we introduced sort of what I would say as in-progress highlights. An in-progress highlight is a highlight that happens or is a piece of content that we're pushing up while a game is going on. And it was great. And we were able to get the content out there in about 10 minutes to 15 minutes or so. And when we released that year, we saw our consumption of that content just doubled from the year before. We hadn't really had that content out there. We've had the content out there in a post-game scenario, and we moved that content up, and the consumption went up. And we continue to keep that consumption through the course of the day. And it was fantastic.

And going through the next year, we said, "Well, wait a second, how can we make this even better? Can we make this faster?" So we challenged ourselves and we came up with this concept of -- we want a 30-second highlight. Get a highlight out there in 30 seconds after the game has happened. We worked very closely with the Akamai team to enable that. And we were able to do that. We rolled that out in 2010, and it's been fantastic. And what we've seen though is that the time between that 30 seconds and 10 minutes, there's a whole new section of people there who are interested in watching that content. So not only did we move up the consumption of the content, consumption grew again. And I think this really crystallized when we actually relaunched our Fantasy game in 2010 as well. When we asked our Fantasy users, we wanted to put this video content in the game and when do you want it? Well, they say, "We want it quickly. We want it as fast as we can. We don't want it in 10 minutes, we've already seen it in 10 minutes. But we want it now. We want it as quickly as we possibly can." And so we've opened up this -- introduced this new aspect of sort of unlocked this new sort of time area where people are just dying and grabbing all these content. And it's been great for our business.

Additionally, I think, as Bob and others have been saying today, it's sort of as the world is becoming much more, I guess, platform and device sort of proliferating out other, we want our content to be in many different places as well. And we need to have that content customized for the device that you're looking at. And we've been utilizing the suite of services from Akamai to make sure that happens. So it's important for us to make sure that where your content is. Whether it's on a tablet or whether it's on -- or whether your viewing the content in Canada or in the U.K., we can be serving specific bytes of content to those individuals, and to make sure that the content is customized for those individuals. Because it's really important for us -- as we see about growing our audience -- how we can grow our audience is serving the content that they need into the devices and to the platforms that they're using today. I think we think about it traditionally or sort of what we're trying to do is create more content, grow the consumption of that content across more platforms and then really high-quality. And today, Akamai enables all of that for us.

So that was pretty much all I had to say, and thank you, very much.

Robert W. Hughes

Thank you, Noah. We're thrilled to be supporting an organization like the NFL, and it's really one of the things that makes us better is this 4,000 enterprise customer base that we have. Just to give you a quick snapshot of who that is, and we support the leading companies across all our key target industries. Akamai serves today more web content than anyone else in the world. And the reason they trust us again is because they know to be successful online, the challenges go far beyond what a traditional CDN can provide. And they know that Akamai can deliver the solution because we built the solutions with the endgame in mind, thinking about these whole entire enterprise, whole businesses online. So it's no longer a video site, it's no longer just speed it up. Speed it up, secure it, mobilize it, analyze it, explain what's going on. And so you could see the stats, they're quite impressive but we have just a tremendous amount of work to do there.

What I want to do now is give you another good example of a big enterprise and how they're starting to use Akamai. We take Hertz. Hertz is now leveraging Akamai, so a one-stop shop for everything that goes online. Hertz is one of the largest car rental brands in the world. They've 138 domains. They operate in over -- now 140 countries, not surprisingly online bookings have become a huge part of their business. It's now critical for its customers do be able to transact online in a fast, secure and reliable manner from any device. It's what happens as you get off the plane, you access your iPhone and you could book, et cetera.

Like most customers, our relationship with Hertz has expanded over time. They initially came to us 3 years ago to optimize the performance of their website with DSA, Dynamic Site Accelerator. They then added to their online business in rolling out the Hertz on Demand and Hertz Rent2Buy business lines, both of which have done extremely well for Hertz. Critical to the success of these initiatives was a quality experience on any device and meeting the compliance standards for secure transactions, as Tom talked about PCI compliance. So with our mobile reader app solution, it helps manage performance and the appearance on multiple devices. So this is critical. You got a big enterprise, there's no monkeying around. The brand is going to appear fast and well represented on that device. On top of that, they added our Edge Tokenization solution to secure these online transactions. This solution saved them over $2 million in PCI compliance costs. Monthly recurring revenue in this enterprise increased from 30% -- increased 30% in its first 3 years. It's projected to grow at another 40% in 2012, with the addition of Cloud Infrastructure Services. So these are the B2B applications that are inside of the enterprise now that they used Akamai. They started on the front end. Now everything is internal, and there are security solutions. Because again, this is -- anybody doing business online suddenly becomes a prime target for what reason has Tom pointed out, we never quite know. This sort of growth is consistent with what we're seeing across our customer base. As customers evolve their business online, they invest more and more into the platform.

Which brings me to our plans for how do we -- our go-to-market going forward, how we continue to expand? So it's really, as we go forward, it's focused on 2 key areas. One, solution innovation, bringing new solutions to market. As Tom presented earlier, we have a whole suite of new products coming out in 2012 that we're bringing to market. In these particular areas of cloud, mobile -- jumped one slide there, backup -- we have a suite of new services coming out in 2012 in these core areas that we have talked about. Some of them are upgraded, some of them are whole new service lines. But the goal is to really drive further growth within the installed base and into new customers. Because some are going to enter at different points. Some might enter through that Riverbed product, some might enter through the website, some might enter through the Akamai platform for video.

An interesting stat here as well, 78% of our customers have purchased at least 1 cloud infrastructure service, less than half have purchased 2. So there's a tremendous upsell opportunity across this enterprise customer base to take these new products that are coming out, some of the existing products, as well as walk across into new opportunities. It's one of the areas when we talked about the go-to-market going forward that we're really directing the sales and services back into these enterprises to sell new -- as well as New Year's passed. They've talked about going wider. We've got opportunities on both sides, thanks to this tremendous push on innovation coming up.

We're also innovating our portfolio by building out the solution partner ecosystem. This is a huge change for Akamai. This is into reseller business. We're calling this the solution partner ecosystem and the numbers will be reflected in the overall channel numbers. If you watched the announcements, we've announced several technology partnerships this year which enable us to bring the most holistic solutions to market across the whole product line. We're partnering with the technology leaders like RackSpace, Riverbed, IBM, to embed our software into their cloud and their hybrid cloud offerings. We're also working with companies like Netbiscuits and Ericsson to expand our mobile offerings.

And we continue to develop our media partnerships to simplify workflows. So what does this mean? If you step back, this is a huge investment in sales capacity that doesn't require the build-out of the direct sales force and it dramatically increases the pace in which new solutions come out and it gets us into the account. As we talked about, one of the key things for Akamai to grow is once we get in, they understand the platform, they start to walk across. This is the first time you're seeing Akamai really open up that platform. And so we're going to put our products inside of other people's products. And we're letting our customers do more and more on that platform and the solutions are built on top of that same software platform, which gives us tremendous scale.

And the second area that we're investing in is -- as part of the go-to-market strategy is global expansion. We've continued to grow our operations rapidly outside of the United States. We're now open for business in 39 markets, with plans to open another 8 in 2012. Last year, we announced -- we opened our headquarters in Zürich, Switzerland. This year, we opened our APAC headquarters in Singapore. The APAC region, outside of Japan, has been the fastest-growing region of our business. We expect that to continue going forward. And the last few years, we've been targeting most of our investment outside of the United States. The growth we're seeing this year, clearly has been impacted by some of the macro headwinds in the bigger economies. Some of our investments have taken a bit longer to bear fruit, but I'm very confident, we're very confident there's a tremendous opportunity outside of the U.S. Again think about the population, think about the pervasiveness of the Internet, the globally distributed workforce, all the trends that we've talked about are not unique. So it's in North America. So we're going to continue to build out, and you should see tremendous improvement in these results. We're not happy where we've been, some of it we can't control, some of it we can. It is taking a little while the ramp and to localize our product offerings.

In addition to building out technology and solution partners, to accelerate that innovation, we're also investing in our reseller partners to expand our reach. We've grown our channel revenue this year, 15%, so faster than the core business. We've added 34 new partners in EMEA and APAC in the past 6 months. And we're really excited about the opportunity that lies ahead for us. Again so when I say increased sales capacity, how we're doing it, direct new offices, existing reseller channels, technology solution partners to complete the solution and bring theirs and again seed Akamai deeper into these enterprises.

So if you think everything that we've talked about, it really boils down to 3 key points. First, the market's ripe for our solutions. The hyper-connected world is in its infancy, the markets rapidly evolving in the Internet, it's really the playing field for the business. And I hate to use the term playing field, because it is business -- it's the whole enterprise is now moving online which just represents tremendous opportunity. I get a lot of questions they say, "Are you worried about the move to cloud?" "No, the best thing in the world that can happen is to move to cloud." Cloud is the future, next-generation of hosting that's going on in the world. All this money is being made in hardware sold in apps are moving. If you take away one thing from today, think about the workforce becoming more distributed, companies becoming more global and there's going to be more devices. Ultimately, the fundamental problem, and no cloud can address this, is its in network optimization problem. That's why every cloud wants to partner with Akamai. That's why every enterprises are saying, "I'm moving this app, I need you attached to the cloud." Because again, you've heard the story early on the decade with the hosting companies 1 data center, 2 data centers, 3 date centers, 4, 15, 20 are the most in the world. They're the biggest, baddest, most secure, but if you heard Tom and Bobby talk about it, it's a network optimization problem. It's a security problem. No amount of data center is going to solve this. And so it's really ripe. And so we're very excited and what we're seeing now is as they move to the cloud right along with it is, "Wait a minute, I just spent all this money, it's working, it's doing better than it was. There's nothing wrong with the cloud, the data center is more effective." But the users are on the move. Think of yourselves as you traveled here. You probably were at home, you might have been on your files connection, you left, suddenly on a wireless connection, suddenly you're in a hotel, you're crossing networks, you have your laptop, your phone, et cetera. You want all that experience to be seamless. It is ultimately a network optimization problem, and no one data center or 50 are going to solve that.

Second point. We're very focused on global expansion and innovation. We have a great install base, tremendous amount of greenfield outside the United States. We're going to keep investing there, we need to go faster. And back in the install base. Only 2 -- if we look at the stats, 80% of the customers have bought 1 cloud service, less than half, have bought 2. We've got about 5 more coming out, plus a whole range. Look at the market trends, and you could see why we're so bullish about the opportunity.

And lastly, the third point is, we are a complete solution. We focused on the customer, we thought about it for the whole decade. What are they going to do online? And we're adding that functionality. And so this is a huge change. Again, when we were -- early on, when you picked up a dot-com customer, there was one thing they were doing online. When you focus on the Fortune 1000 and they're talking to you about moving the whole business online, you can't come in with just a performance solution. You can't come in with just a security solution. You cannot come in with just a mobile. You've got to walk in with the entire suite and the platform interface, the portal that we talked about, because you're going to give visibility to the health of the Internet so they know what's going on with the world, you're going to give them help with what's going on with their app, and all they really want to think about at the end of the day is their business. They want to think about: I'm putting this up, what's happening with the user? We have to -- our job is to strip all that complexity out. Do video, do whatever you want to do and we're going to detect what device, we're going to make it perform and we're going to handle all your apps.

And so with that, again, I think we're very uniquely positioned for growth going forward. We do represent the complete solution. There is competition out there, we do respect it, but once our sales team is in there with these partners, with our services team, we differentiate ourselves very, very quickly.

And so with that, I'm going to turn it over to JD.

J. D. Sherman

Thanks, Bob. Good news, I'm the last speaker. So we ran a little bit long, but hopefully, you're finding some value in this day. And I, last week, realized just how important Investor Day is, at least for us, to be able to communicate to you. I was at an investor conference and I brought with me Jim Benson, who is our Senior Vice President of Finance. He sat in the audience. He didn't have his nametag on, so nobody knew he was an Akamai person. And the investor sitting next to him leaned over during my presentation and said, I have no idea what these people do. So that's -- now we're starting to know what we're working with.

And so the objective of today is to help people understand a little bit more of what we do. We talk all the time about what that means in terms of our financials, and now we'll cover that a little bit. But hopefully, you've got a little bit more cover -- color on what we do and the power behind what we do. So I'm going to try to wrap that up a little bit.

You've heard us talk about the industry solutions portfolio. Paul talked a lot about the growth drivers that we see that are driving the growth for Akamai now and we think into the future. I want to put that in context of our business model.

So let me start there. And actually, let me start with the chart that I used way back at the first Investor Day we did in 2006, my first as the CFO. Frankly when I didn't know what we did. But the importance of this chart was to talk about what I considered the power of the Akamai business model, which is, it has one aspect of it which is a services-type aspect or a network-type aspect, building recurring revenues. And as those revenues build, and as the business builds, you get tremendous operating leverage off of that, and that's our content delivery business largely today. And then it has another important aspect, which is software-like. It's software fundamentally. So to develop that software is very low capital expenditures, and it's very highly scalable. You write the software once, you apply it across the solution portfolio, and the industry verticals where we go to market. So I think that's still true today, and even more true as we've developed into a $1 billion-plus business. And the way we have sort of realized that and talk about it today are across the solutions that we take to market, we think about 2 sort of sets of solutions, or sets of businesses. One is content delivery solutions: media delivery, software delivery, object delivery for website. Really, the volume is driven by traffic on the Internet and the fundamental drivers of growth on the Internet, which are more and more today driven by video. And then on top of that, we have our cloud infrastructure solutions, where we're accelerating dynamic content, where we're extending the value of the platform into security and mobile, like you've heard about today, where we provide analytics to customers and custom services and solutions. And so that's the cloud, that's the software-like business.

Now you'll notice the subtlety here. I haven't called it volume and value-add, right? That was the mistake that I made several years ago because I don't have a marketing background to call it that way. This year, I collaborated with Bob's team in global sales and marketing, and I think we've come up with more descriptive names for these types of solutions in these businesses. Because really what we're talking about is a platform that Bobby and Tom described great, which is the fundamental underlying of our entire business. Content delivery solutions that we build on top of that, which are driven by the traffic of the Internet, and basically get us to scale, give us the license for that scale. And then software or cloud infrastructure solutions that's right on top of that.

So with that, let me talk about and give you guys an update on the industry verticals that we talk about quarterly and give a little more detail into that.

So let me start with year-over-year growth. So in 2010, our business grew about 19%. We reached the $1 billion mark. This year, we're going to do, if you use the midpoint of our guidance for Q4, and we're not going to update guidance today, that would be about 12% growth, or just $1.14 billion. I think what's really interesting to look at is the 19% and the 12%. If you look at it by vertical and what's been the profile of that growth. The obvious thing that comes out of this chart, obvious partly because I circled it in orange, but obvious even if I didn't do that, is that the commerce and the enterprise verticals, which are largely where we're driving into the marketplace with our cloud infrastructure solutions, are growing as fast or faster than last year, where we grew 19% as a business. As fast or faster, commerce down a little bit, but I'm using a little license there. In total, as fast or faster. If you look at media and high tech, where we're seeing a bit of the downturn, we'll go through each of those, but that's where we're seeing the impact of slower volume growth, slower traffic growth this year. It's interesting and I'll talk about this when we get into media, but let me talk about it right now.

If you look at the growth of our media vertical, over the last 4 years, we went plus 24%, minus 4%, I think, plus 20%, and this year, it will be 7% or 8%. So it's bounced around, driven by the small rates of change in the volume growth. So volume is still growing tremendously, but we tend to see is our inflection points where you get a burst of traffic. I think last year, we saw the average bit, the stream quality go up and that really gave traffic growth on video a real kickstart. But that -- we've seen that bounce up and down. But fundamentally, particularly in our cloud infrastructure solutions, we're seeing great growth continuing.

So let's go through the verticals. If I can get this to move, let's see. I know I have more charts. Yes.

So the commerce vertical. This is the vertical primarily for our online retailers and other businesses that are conducting commerce online. Today, 90% of that business is driven by our cloud infrastructure solutions. The traffic of a commerce website is really almost de minimis, when you're talking about everything that goes on a commerce website. Certainly, from the way we sell our products, there's a very small portion of our revenue comes from the traffic. And actually you can see over the years, that's declined. 88%, almost 90% of our business in the commerce vertical is from the cloud infrastructure solutions, and they're growing obviously faster than the average 22% this year. And the financial model is very much like a software model as a result of the mix of businesses here. Very high gross margins in the 80s, EBITDA margins in the 50s, again, higher than our average. And I think another point, and here you can see the realization of what Bob and Tom and the team have talked about, look at the ARPUs for our average customer in the commerce vertical. They're up 15%. So we're growing 20% this year, 15% growth from ARPU. So really critical here is up-selling these customers. As they grow, up-selling them to DSA, up-selling them or cross-selling them on our entire portfolio of solutions. And that's really what's driven the growth for our commerce vertical in 2011.

Moving onto the Enterprise vertical. Very similar dynamics. Here we're about almost 80% of our revenue, and growing even faster than the commerce vertical. Again, the content delivery part of this business is relatively small. The economics, very similar, even slightly higher on the gross profit margin. Gross margin's 85% to 90% in this vertical. The EBITDA is still right around where it is in the commerce vertical, where a lot of our investments are focused particularly on products and solutions that we're driving into the enterprise vertical. So a little bit heavier investment. And again, you see the ARPU, which is lower and smaller than with our commerce customers where it's $23,000 a year. But $14,000 a year is up 17%. And to me, that's a really powerful stat, because this is also one of the verticals where our new customer ads have been significant this year. And as we've talked about, I've talked about with many of you, most of the time when a new customer comes online, they're well below our average ARPU for the business. So to get 17% average ARPU growth here, even as we add new customers, I think that shows the power of having a total solution and a total platform. And I think it also shows the potential to drive growth in the enterprise.

Looking at the media vertical. The media vertical is obviously dominated by media delivery. One interesting point I'd make on the media delivery is if you look at the growth of that blue, the media -- the volume, the content delivery solutions, it's driven by video. It's the one place where we are seeing growth in our content delivery solutions portfolio. Media, video -- revenue from video delivery is growing just over 10% this year. So we're seeing software delivery decline a bit, and you'll see that in the high tech vertical. We're certainly seeing object delivery from commerce websites and other websites decline. The video is driving growth. And the financial model for our media business, obviously you can see the ARPUs are much larger, and the ARPUs have actually come down this year, right? We've grown media 8% while the ARPUs have come down, so we've added more new customers to the media vertical, but because of a slower rate of volume growth, we've seen the average ARPU come down and the margins are much more like we see on the volume side of the business with the gross margins in the 50s, but less investments, so EBITDA margin is still 35% to 40%.

Looking at the High Tech vertical and we've talked about this. We basically have a tale of 2 worlds, almost a microcosm of Akamai here. We have a content delivery business that's primarily driven by software downloads. And you can see 2009, 2010, 2011, the software download business has declined. Basically, there's -- all software gets delivered on the Internet today. That's obviously a bit of an overstatement. But what we saw again is from 2006 to 2009, where that shift was happening. We're pretty much at the turning point of that shift. Software volumes are still growing, but they're benefiting from the price declines that are being driven on the video side. So I would expect to actually to see that blue part continue to decline slightly over time, when I would also expect to see the orange part of this bar, the cloud infrastructure, where we're selling Software-as-a-Service customers on our cloud infrastructure solutions to continue to grow. And this was a year where we've crossed over the 50-50 point in terms of our revenue in this vertical. And the 2 are basically battling it out to flat year-over-year in this vertical. Economics here is sort of a mix between the 2 businesses. We see gross margins that are 70% plus. EBITDA, because we're not really investing a lot in new products, and also get a lot of leverage and sales model here, are very good, better than average, and large ARPUs because these tend to be larger customers and the ARPUs are down slightly this year.

So let's talk about the 2 businesses now. And it's a bit of an oversimplification of our business as I've talked about because they both ride on the same platform, but I think it's helpful to think about the economics of these businesses somewhat separately. We certainly do, and I'll talk to you about how we -- what we try to achieve with the business models for content delivery in our cloud infrastructure solutions.

So content delivery service is about 40% of our business. This year, it's growing low single digits. Our objective with the content delivery service is going to be driven by traffic, right? The growth of this business is going to be driven by traffic, and that traffic growth is going to be driven by online video adoption. So we want to grow as fast or faster than the traffic growth allows us, given that prices are going to come down. Is that 0% growth? Some years, it might be. Is it 20% growth? Some years, it might be. This is the wildcard in terms of what we've seen in growth for our business. Again, going back to what's happened with our media vertical, which is largely driven by volumes. We have seen the growth rates bounce up and down. But we want to grow as fast or faster than traffic on the Internet allows us to grow in this space. And we want to return our cost to capital or better. It's not going to have software margins. It's going to be a bit more capital intense. But if we can return our cost to capital for this business, that effectively funds the scale of our overall business and it's a very attractive model. So the financial model for this part of our business is sort of, to put it in a ballpark, 60% cash gross margins, 35% EBITDA margin, and 20% to 25% CapEx as a percentage of revenue.

Look at the other side of that pie, if you will, cloud infrastructure services. This is really like a software business. And the drivers in this business are what you heard about today. We want to grow this business as fast or faster than transactions move online and applications move online. We think that's a very -- we think those trends are very rapid. Paul talked about cloud computing adoption and spin on cloud computing being a 20% CAGR over the next 4 or 5 years, talked about online commerce growing in low-to-mid-teens over the next 4 to 5 years. We think we can grow as fast or faster than those external benchmarks and we think we've demonstrated our ability to do that over the last few years. And this business is going to have software economics. Gross profit margins are going to be in the range of 90%. EBITDA margins are going to be well above our average, 55% plus or minus. And CapEx, 10%. And really what that CapEx is, is capitalizing software development. There's really not much network investment. We don't need to add a lot of servers or a lot of storage to support these cloud infrastructure solutions. In some sense, they get a free ride, at least economically, on the backs of our content delivery business.

So when you blend those 2 things together -- you guys have seen this chart before. This is the same chart we've put up 4 or 5 years ago on our long-term financial model. It's obviously mix-dependent, but if you do the math and blend those 2 businesses together, we're confident that we can drive cash gross margins in the high 70s, EBITDA margins in the mid 40s, and spend about 13% to 16% of our revenue on capital expenditures. The one minor modification I made to this chain -- to this chart from last year, as you will probably notice is that the GAAP tax rate, we took up a point or 2 based on the mix of our businesses internationally and the rate of change of that. Last year, I had 30% to 35%. Here now, I have 32% to 36%. But we think fundamentally, those 2 pieces of our model, if we can execute on growing with the Internet on the content delivery business and returning our cost to capital, and then we can build on the products and solutions that you saw today, and our strong position in the marketplace, to drive our cloud infrastructure businesses, we are confident in this as a long-term model.

That model is a pretty powerful model, we think. And the return on invested capital for the model, in total, has been good -- very good and growing from 20% plus a few years ago to in the 25%, 26% range today. And we're obviously very focused on excellent return on the invested capital of our business. It also generates pretty significant cash flows. So this chart shows from the beginning of the year to where we ended Q3. In terms of the elements of our balance sheet, if you will, from a cash perspective, we generated just over $300 million. We're well on track to generate over $400 million of cash from operations this year. We'll spend a portion of that roughly something like 1/3 of that on capital expenditures for the network if we stay in the range of 13% to 16%. And then we'll use some of the cash from operations to buy back shares. That gives us a pretty significant amount of balance sheet flexibility with $1.2 billion in cash. Some of that flexibility, we actually used last quarter to make what I think was a pretty smart investment or a pretty good investment in Akamai stock and accelerated our share repurchase a bit last quarter. But I think our fundamental objective with the balance sheet is to make sure that we have the flexibility to make some of the acquisitions that can help us drive our product portfolio and expand the relationships that we have with our customers. And I think we are in a very good position to do that, and I think this model allows us to continue to build on that, drive cash from operations that more than fund our CapEx requirements and allow us to also be at least buying back shares to offset the dilution from equity compensation.

So I went kind of fast because you guys hear from me all the time. We'll certainly have a chance to answer questions. I'll invite the team back up to answer questions. I'll also just point out that we'll be around a little bit afterwards so that we can mingle and answer any specific questions. There are kiosks over here with some demonstrations of our products and you can ask the technical guys who actually know what they're talking about a little bit more about the products. There are also some downstairs. There's going to be food upstairs and downstairs, so make sure you check out sort of both rooms. The charts that we've shown you today, we'll make available on the Investor Relations section of our website so that you don't have to -- I see a lot of you guys typing very rapidly and scribbling down. We'll make the charts available.

So I just want to leave you sort of with my CFO summary of the overall themes of the day, which is we have a very unique asset here in terms of the Akamai Intelligent Platform. It's the asset that underlies all of our solutions and what enables us to have the scale that we have and enabled the breadth of our portfolio. On top of that, we have a business delivering content over the Internet. It's bigger than anybody, it's at a better scale, it has industry-leading performance in cost and it's being driven by video adoption. And then on top of that, we built a software business that we call Cloud Infrastructure Services, and that's a software business. Software-as-a-Service, it enables business models online, and cloud adoption is going to be a key driver for the growth of that business.

So with that, let me ask Bobby, Bob and Paul to come back up, and we will take any other questions that you have.

F. Thomson Leighton

So point to the mics and then...

J. D. Sherman

Yes, we'll get the microphones so that everybody can hear on the webcast. Mark, do you want to start?

Unknown Analyst

Wanted to wrap together a competition and a pricing question. So these 2 segments, maybe Bob, where do you see greater competition? If you think about the business, CDS and CIS, where do you see greater competition? Where are you more likely to see pricing pressure? And can you talk about where Amazon, with AWS, fits into that competitive landscape? They're doing $1 billion in revenue in their cloud business, which is -- I know there's a lot of differences, but there's a lot of similarities, too. How do you think about that? How do you see them in the field?

Robert W. Hughes

Sure. So pricing competition, more prevalent on the CDN side of where -- where we've talked about, that's going to continue. We expect it -- we've seen it for years. We don't think it's going to stop and we're going to keep trying to drive the cost down there as well. So that's by far where we see it. That's where we see the bulk of the competition coming in. In terms of the cloud, again, they're all great partners. They're great customers. You see how we're embedding inside. Again, what's happening in that data center is not what we do, but a lot -- I think Carl [ph] has done a great job of making it easy to get the app -- move their app up into the cloud. It's great. They've saved -- the Enterprise saves money, it becomes distributed. As we've talked about in the past, one of our biggest hurdles to overcome has been do-it-yourself or get out of this mindset of we've got to build the data center. All the clouds are actually helping us. They're saying, "Hey, distribute this app." Once it's distributed, the need for our services actually continues to grow. So Step 1 is they're getting it out there for a very good reason. They're going to save money, it's more efficient. But if that app scales, it has to perform. At that moment, again, it's a great partnership with the clouds, because now it’s a network optimization problem. So there's great -- I don't want to slight the clouds in any way. They've done some great innovation, they're doing great work inside the data center. But the problem is, the user is going to cross multiple networks around the world. There's an additional huge market that we play in, which is the network optimization problem.

F. Thomson Leighton

And I think the piece of players who are providing hosting as a cloud, which is really hosting by the drink. At some fundamental level, it's what's old is new again, has opened up new market opportunities. In some cases like RackSpace, it's an Akamai upsell product, brand name, et cetera. In other cases, it may be white label opportunities. In other cases, we won't have all the relationships and we simply go into their customers and fix performance problems. We've seen over and over again, people who start in these cloud infrastructure places, basically with a small world product or project on a credit card. And as they grow and scale matters -- but frankly, actually, cost and performance matter -- they migrate to us and we can either sell in partnership or sometimes we sell on our own as an add-on, on top. And where we can get embedded everywhere, we'll go with that model and where we have to go, say, actually we can make this better and more cost effective on top, we'll sell that way as well. Because I think you're going to see direct hosting companies moving this way, you going to certainly see the telcoms moving this way, and then you're going to see new entrants like the one you mentioned or many others coming in. We'll try to partner where we can; and where we can't strike a deal that we think is fair, we'll sell on top.

Unknown Analyst

Question on the new licensing solutions. How do plan on going to market with that? Would that be a direct sale, I assume? And then how will the revenue models work with those solutions?

F. Thomson Leighton

I think it's a new model, so it's going to have to start direct. And when we started this company over a decade ago, nobody really knew what we were doing. The people who, often who listened to us said well that's interesting but impossible. So we had to sell direct. We had to go evangelize and prove it. I think the efforts that Bob -- we talked about, we're going to have to start there. But clearly, the leverage is through partnership. We're already seeing that opportunity in mobile with what we're doing with Ericsson. I think some of the other network operator solutions that Bobby described today, we're going to sell direct in the beginning. The ideal for us would be to embed in the top 3 or 4 global players and let them make that part of their solutions.

Unknown Analyst

[indiscernible]

F. Thomson Leighton

I think it's going to differ. I think in some cases, we manage, and we're going to get paid. In other cases, we'll allow people to manage and we'll get paid something. In some cases, on the license model effectively, we're going to sell them capacity in one form or another, by processor, by hits, by deliveries, by streams, and we'll get paid that way. It won't be on the hardware model, though. Our goal is not to be a hardware business.

Unknown Analyst

On the network operator solutions business. I was just curious, how much revenue -- how you want to define that -- are you currently getting from network operators today? My sense is that you're obviously very dependent on them to allow them to let you put your servers in their networks. And it now seems like you're going to try and find ways to actually start to monetize that relationship. In addition, what gives you the confidence that they're going to be willing to pay you for some of the things that you want to sell them?

F. Thomson Leighton

Well I think a couple of things. I think I would turn your question around and say they're very dependent on us to brings them these solutions that we've brought for over a decade. And I think the point Bobby made is, despite all of the noise that we hear about how we're less popular, we're more popular than ever. In fact, if we got turned off, if we just got really angry and went -- Tom went back and actually took the key he's got, the internet would just stop. Just stop, right? If we're delivering 30% of the Web some days, that traffic simply cannot go anywhere else. So the networks are actually I think, we have a very close relationship. What we offload into their network and take off their backbones or deep into their networks as Bobby said, simply can't go anywhere. And if they had to go provision for it, it would break all the economic models. So I actually look at it a different way, which isn't, "Oh my god, it's great they want to talk to us." I think they think it's great that we show up. And we hear that over and over again. We do think and we see them spending particularly around the video problem, enormous sums of money on solutions that often actually don't really provide much value, and they've been asking us for this. There will be a push-me-pull-you. Well shouldn't you do it for free because we're partners? And we'll say you should pay us way too much money and we'll have to negotiate and there will be other solutions and we'll live in that competitive environment. But I think there's a great opportunity over the next 5 years to bring them new value on stuff that isn't our business. I think what's our business is, kind of say, "Well, we're helping you a lot. Why should we pay anything?" And we'll say, "Maybe pay something." I think where it's not our traffic, if you will, we certainly have no reason to do that for free, and they're spending a lot of money making it better. So I think there's a really interesting opportunity to partner more broadly, more deeply and to get paid for some new things that we think we have unique capabilities of doing. But frankly, our relationships with them, this isn't fly-by-night, we've been doing this for 13 years. And we've left virtually no networks, and we just keep adding as the world has changed. So I think there's just this huge opportunity for us there.

Robert Blumofe

With the volume of traffic that we carry and the advanced solutions, we're more relevant to the networks than we've ever been.

Unknown Analyst

Could you comment on some of the partnerships? RackSpace, Ericsson, IBM, Riverbed, how those have tracked in terms of your technology expectations so far? Also, could you comment on progress in terms of developing sales models and your expectations for revenue contributions, specifically from Ericsson and Riverbed?

Robert W. Hughes

Sure. So I will put -- Ericsson is actually -- So in the other models, this is a huge change for us. We've had to resell our business for a while and companies like IBM and Verizon have been great reselling partners for a very long time. This notion that we're opening up the platform, and letting other people develop on that platform or taking our technology inside it, it's a huge change. We've been really spending a lot of time on the last couple of years to do it. So we do have very high expectations for where this goes, because ultimately, it's a huge sales force multiplier and it seeds our technology into the account. So we've -- in case of RackSpace, their service is up. And it's a matter of attach rate. So you basically go online, you order it. And we can attach basic, very basic delivery services. One of the challenges we do face as a company is the term CDN is just abused out there. There's lots of different flavors. So the basic delivery services, what we will attach to the cloud, the more advanced services and some of that embedding work that's all in Riverbed, that's a completely different portfolio. But again, it's a huge sales force multiplier for us and sold through them. We're in beta right now, so that comes out in Q1, so we're going to speak more to that going forward. But I'm very, very excited about it. But it is a significant shift from where we've been, predominantly direct or sell with. This is a model with resellers, we'll continue to sell with. With these solution partners, they're taking it to market as their product.

F. Thomson Leighton

We've never embedded technology that I can recall in a third-party product that someone else could just sell and turn on like the Riverbed solution that's out there or the Ericsson solution that's really more an alpha, but it's actually being trialed today. But the permanent solution is commercial public beta traffic oversubscribed this quarter running. And other people will sell that. Of course, we have to assist and explain it. That's a very different model for us.

Robert W. Hughes

But it's been -- the key there is in built -- the technology is built from beginning to be put there. And so now, one of the folks here in the room asked me during the break, do you expect more? Absolutely. This is a key part of our strategy going forward is it can be Riverbed, it can be -- just think of all the different apps that are out there and the traction, we're absolutely looking at that. We want to get this one right, but there's no stopping there with this model. Because again, the nice thing is they're asking for it. What we had to do and it took us a lot to do, is change the way we built our products and open up the platform. And that took us a lot to do it, but we're very excited that they're largely going to sell this product, it's a Riverbed product, they could support it, troubleshoot, sell it. It's there and again the customer will know Akamai is inside there, and then we can walk in and sell the rest of our services.

Unknown Analyst

Can you give us a little bit more color on how you divide the cost between the 2 sides of the business? Because for example on the CapEx side, is at 25% on the CDN and 10% on the cloud infrastructure. Yet cloud infrastructure leverages a significant portion of the CDN infrastructure.

J. D. Sherman

So the way we think about it is basically, the amount of traffic that's driven by each side of the business. A huge fraction, probably 90% plus of the traffic, is on the content delivery side of the business. You're on a very fair point though, which is, it will be difficult for a business if some business just decided they -- I don't want to do the traffic piece of that. To build out that massively-scaled, massively-distributed platform would be more expensive. So you get effectively somewhat of a free ride. So there's a bit of unfairness there in terms of the way we allocate the cost. But again, what we want to do is we want to think about this massively-scaled platform, delivering content on that platform and returning the cost to capital of the platform based on our content delivery business and then snapping on, if you will, this software business and getting great leverage on the model that way.

F. Thomson Leighton

But to give you some insight, JD only allows the business unit general managers to negotiate their allocation one time a year.

J. D. Sherman

Yes, that's right.

F. Thomson Leighton

Mike?

Unknown Analyst

If I can squeeze in 2 questions. One, real quick, I think you mentioned that there'll be new product rollouts next year. I think someone said 5, so what are they? And then my other question is on margins. You're below the target margins, the long-term target margins right now, and I understand how the mix shift towards cloud will drive margins up over time, since it's just faster growing. And yet, this year, margins went down even though cloud did grow faster. So what were the dynamics in hurting margins this year, even with the positive mix shift? And are those over? Have we worked through that and that starts to turn around?

J. D. Sherman

Let me -- so I'll answer that question, then we'll go to the other one. So, yes, it will be about 44% EBITDA margin in Q4, so a bit below our long-term model. What we saw this year was -- well first going back, we put that long-term model out and then we've actually blew past it up to like 49% EBITDA margins. And as late -- as recently as the end of last year, revenue growth took off and actually outstripped our investment. This year, we've kept the investment levels growing, particularly on the cloud infrastructure solution side, even as we saw the volume growth slow and the margins decline a bit on the content delivery solutions side. It's always tempting, especially if you're me, to want to make sure you stay within the model, even if it means cutting investment for the future because it's nice to talk about how you're in the long-term model. We wanted to make sure that in a year where there was a bit more pressure on the content delivery side of the business, because volumes had moderated and we've seen some gross profit margin decline on that side of the business, we want to make sure we're still investing to drive the growth in the cloud infrastructure business. And then, I don't know if Paul or...

F. Thomson Leighton

Yes, I think Bob went through them but you could list them off again.

Robert W. Hughes

From the part -- we're certainly not going to launch the products here. It was mainly in the categories that we talked about. So you'll -- for sure, we've talked about the one Riverbed product that comes out in Q1, that's the hybrid cloud product. It's a suite of security solutions that come out in Q1, and then advancements in the -- that umbrella of cloud infrastructure services, particularly on next-generations of DSA, our application performance solution in mobile. But we're not going to launch the names here today.

F. Thomson Leighton

All right. And then there's some on the HD --

Robert W. Hughes

But those -- I think those are really the categories. The next generation of those solutions, plus the additional hybrid cloud.

F. Thomson Leighton

And some enhancements and new stuff on the media side as well.

Unknown Analyst

I think Bobby, you touched a little bit on your peer-to-peer technologies and we looked at someone in Russia with the client software. Question is really, how much are you utilizing that for content delivery today? And is -- are you picking up content from resident on client to distribute to another client? Or how is that working?

F. Thomson Leighton

I think Bobby can describe the process. I consider proprietary, the percentage. You saw on Tom's display, the amount of offload that those clients happen to be serving. We won't give you a network-wide number, although it's significant to capacity and cog.

Robert Blumofe

Yes. And it's hard -- the clients is really another way to extend the network and have it work more efficiently, more reliably, and at greater scale, at a lower price point. And it really is sort of a logical extension of the Akamai network. It's very hard, by the way, to make that kind of delivery capability work without a connection to a core platform. Pure client delivery basically doesn't work. Client-assisted delivery, connecting to a platform, that works. And that's what we've been demonstrating. That's what is working, and I expect that to continue to scale and be a big part of our solutions going forward.

F. Thomson Leighton

And if you think about--

J. D. Sherman

I had a guy from Russia calling me asking to send him 10-bit coins for using his server.

F. Thomson Leighton

If you think about how do we keep the number of servers that we have to own as small as possible, using all those other processors and connection points as a piece of it, so growing that over time is a really important piece of not just capacity, but economics.

Unknown Analyst

Does it bring up rights issues?

F. Thomson Leighton

Well, the customer or the end-user accepts it, so there's no issue there. They accept it. And the key point, not to be messing with what Bobby said, is it connects to the backplane. So we can authenticate the user, we can record it, we can make sure it's conforming with all of the other rights that our customer would have or be restricted by as if it was being served right off their website, off the origin. So the key is not that it's running in the wild, the key is that it talks to a control plane. And so all the rights issues, all the reporting, all the monetization, all the security, is wrapped in that offer.

Unknown Analyst

Sort of an industry trend question. When you think about the Fortune 500 and the private networks that they've engineered on a global basis with their various telco partners that they've had over time, and then you're talking about moving to hybrid cloud strategy. As you're having conversations with these IT managers, to what extent are they saying, it was that the VMware for a long time was considered sort of a test platform, you used this as test, it was a nice little test, that you look at stuff with that. But now that it's become even more commercialized, at what point do you think there could be a dismantling of these private networks and moving over to Akamai as to be the core source for what they do across? Or is there going to be this period of time where they are straddling the 2 for 2 different --

F. Thomson Leighton

I think the long-term trend is going to be hybridization. I don't think people are going to shut all their data centers down. That just won't happen. But they're going to say, "I want to stop building stuff that I own and drive it up to peak as much as possible, and connect in a hybrid world. And I think what they're going to look at is, in the same that they've virtualized, we've all virtualized hardware in the data center, the next big -- and starting to do that with storage. The next big piece is virtualizing your network. Not nailing up dedicated point-to-point connectivity because you're going to have too many points and it's too expensive even if you get it provisioned and installed. And what we start to hear from CIOs at Fortune 500 companies is "I'm doing much better on my hardware utilization, my space utilization, how about now, my dedicated networking utilization where the utilization is really low. And how do I go attack these enormous telecom bills? Because I still have a dedicated networking and now I have the Internet and all those locations and then I have the Internet and all the other locations home, et cetera that I'm not in. Can I start ripping out the dedicated networking, just expand my standard Internet a little bit. Can't somebody give me security performance and accountability?" And that's where we want to raise our hand, so we think that's a huge opportunity coming in our direction, but it will be from virtualizing the network layer, effectively making sort of a fourth layer of the -- as-a-software piece, the network-as-a-software piece, and virtualizing that as well. But I don't think you're going to see too many people, certainly Fortune 500 companies saying "I have no infrastructure anymore."

Robert W. Hughes

But I think that conversation -- they definitely come in and say, "Take me out 5 years, what could happen." as Paul said a typical conversation goes. Because there's not a lot of things we don't do, right? That network. But then as they give me the new locations, they give me this. Well what if -- how would this go? So the conversation comes up, and as Paul said, it's going to be a -- new will go to the Internet, the old will stay for a while. There'll be continued -- we don't send a sales force to target and pay less replacement. Their price just keeps going down every year. It's not where we going. Strategically, the conversations do come up saying what if? So I think it will be a straddle -- I think we'll use the word straddle for a while.

Unknown Analyst

So just 2 follow ups. One is having to do with VMware being embedded in the Riverbed services platform. So thinking about a company that's going to expand into a new shopping mall or a retail company. Instead of having to put all of the physical assets into that incremental location, they put in a Riverbed steelhead that connects back and then they can provision a virtual image to that specific retail location and it's a lot more seamless and clean than having to have all that stuff being maintained on-site. And you see that kind of conversation driving the simplicity of really leveraging the virtual view that you get to be able to then provision all that stuff from the centralized data location as opposed to managing all those things at -- or having to --

Paul L. Sagan

And this is beginning to happen. We've seen some retailers, some franchise businesses that have thousands of locations start to do that. We've had some models where we actually deployed in the early days. And a single Akamai server in those places is more efficient than what they used to do but not terribly efficient. Now if you think about that idea just Internet connectivity to each of the retail locations to a pretty dumb thin device with steelhead or something else there, we think that's just an enormous opportunity. Because their -- to their store and someone else's mall, they don't want to bring private dedicated networks. But on the other hand, they may have to update software our offer Wi-Fi in the shop as well and all sorts of things. And that problem of that Dana showed you earlier of choking the pipe with relatively small users and having this wildly different experience isn't acceptable for example if you're a retailer point consumerism. So we think there are lots of opportunities like that but key will be embedding technology and letting other people do some of the selling lifting.

Unknown Analyst

I have 1 off-the-wall question for you. Hope you guys will forgive me for this one. But in my nefarious mind, I'm sitting here thinking isn't there an opportunity for you to go to insurance companies who offer up business interruption insurance and say you really got to have as a checklist item for your people who have -- getting insurance from you, you have to have Akamai protecting your server farms or whatever, however you want to picture the thing. Isn't there an opportunity for you to say that we're just not going to allow the insurance companies say, "We're not going to allow you to have business interruption insurance...

Paul L. Sagan

That's definitely a great marketing idea and we'll take that one...

Robert Blumofe

We thought we -- it's a great idea. We thought about it. Again security for us, I can't stress enough but we focus on the enterprise. We've been in this game for a while. We are now coming out. We're very thoughtful about how we take this thing to market. And you're right, at some point, that would be great. It is, a lot of customers look at the end of the day, by using this I mean, it's what they bought, right? They bought insurance. Yes we do offer it a little bit like that, but it's -- there's a lot of tools and things we're developing in partnership with these enterprises, and I think once we get to a certain point, it would be a great conversation to have.

J. D. Sherman

We'll take one more question if there is one more out there, go ahead.

Unknown Analyst

Just a quick question on CDN CapEx and how to think about it. Because you have a dynamic where volumes grow at 0% to 20% and, 30%, pick a number each year.

Paul L. Sagan

Volumes grow more than that.

Unknown Analyst

So volumes grow x and pricing comes down at y. And CapEx should be driven, I think, correct me if I'm wrong, by volumes as opposed to revenues. So my question is, doesn't that mean that CapEx as a percentage of revenues should go up over time?

J. D. Sherman

That way we avoid that is by being more efficient in terms of how we deploy the CapEx. Some of that, you get automatically, right? Every time a new processor comes out, you go down the Moore's Law curve price performance. Some of it, we get unlike anybody else can because we can -- because our network is software-based. So if we improve the efficiency of our software, we get a benefit beyond what anybody else gets. So think about how we map the Internet, figure out the most efficient way to utilize our network. So for all of those reasons, that's the game. What you just described is the game right? We need to drive our CapEx down just like we drive our bandwidth cost down and our co-location cost, which are very correlated to our server utilization. That drives down our cost and then we can drive down price for our customers. That's the game.

F. Thomson Leighton

A little piece about how Bobby gets paid is how well he does driving software efficiency. It's a huge, huge lever for Bobby and for us. And literally we measure that because a server is not a server in our network.

Unknown Analyst

But co-lo costs have been going up though?

J. D. Sherman

Well co-lo cost, like if you went to co-lo facility and asked: What are my costs this year versus last year?" have been roughly stable, I guess. It depends on -- it's like a real estate market, right? In Manhattan, they've probably been going up. In Jersey City, they tend to be going down. But if you can use your co-location space more efficiently, right? If I can put 1 server in where 2 were, they I need half the rack space, that my co-location cost effectively went down by half.

F. Thomson Leighton

That's another reason why driving the efficiency of a box before you can go into client or offloading into someone else's network is really important.

Well, thank you all for your coming. Please stay for a bite to eat. We appreciate you coming. These folks work really hard. I hope you see what I see which is, best set of customers, best positioning, best people and best technology and innovation products, I think, in the history of the company. So we're excited about the next few years and we appreciate your interest and we look forward to seeing you every quarter at least on that little speakerphone and then seeing in person once a year here in our home or on the road. Thank you very much.

Copyright policy: All transcripts on this site are the copyright of Seeking Alpha. However, we view them as an important resource for bloggers and journalists, and are excited to contribute to the democratization of financial information on the Internet. (Until now investors have had to pay thousands of dollars in subscription fees for transcripts.) So our reproduction policy is as follows: You may quote up to 400 words of any transcript on the condition that you attribute the transcript to Seeking Alpha and either link to the original transcript or to www.SeekingAlpha.com. All other use is prohibited.

THE INFORMATION CONTAINED HERE IS A TEXTUAL REPRESENTATION OF THE APPLICABLE COMPANY'S CONFERENCE CALL, CONFERENCE PRESENTATION OR OTHER AUDIO PRESENTATION, AND WHILE EFFORTS ARE MADE TO PROVIDE AN ACCURATE TRANSCRIPTION, THERE MAY BE MATERIAL ERRORS, OMISSIONS, OR INACCURACIES IN THE REPORTING OF THE SUBSTANCE OF THE AUDIO PRESENTATIONS. IN NO WAY DOES SEEKING ALPHA ASSUME ANY RESPONSIBILITY FOR ANY INVESTMENT OR OTHER DECISIONS MADE BASED UPON THE INFORMATION PROVIDED ON THIS WEB SITE OR IN ANY TRANSCRIPT. USERS ARE ADVISED TO REVIEW THE APPLICABLE COMPANY'S AUDIO PRESENTATION ITSELF AND THE APPLICABLE COMPANY'S SEC FILINGS BEFORE MAKING ANY INVESTMENT OR OTHER DECISIONS.

If you have any additional questions about our online transcripts, please contact us at: transcripts@seekingalpha.com. Thank you!

Source: Akamai Technologies, Inc. - Shareholder/Analyst Call
This Transcript
All Transcripts