Cisco Systems, Inc. (NASDAQ:CSCO)
February 14, 2012 5:00 am ET
Doug Webster -
Robert Pepper -
Gabriel Solomon -
Martin Cave -
David Hendon -
Brian Williamson -
Unknown Executive -
Well, welcome. Welcome. From laptops to tablets and smartphones to such things even as automobiles and refrigerators, it seems that just about everything these days is being connected to the Internet. And more often than not, it's being connected to the mobile Internet.
So greetings from London to our global audience today where we're broadcasting from the Royal Opera House. It's our pleasure to actually be able to present to you from this great city, always on the move, and also, of course, because it's the host of the 2012 London Olympic and Paralympic Games of which Cisco is proud to be a supporter. And today, it's also our venue for the release of the Cisco Visual Networking Index Mobile Forecast spanning the years from 2011 to 2016.
Now today, I'll go over the highlights of the study, actually go talk about some of the main drivers that are taking place. We'll go into some of the factors that are actually helping service providers offload that traffic on to other ones, as well as to talk about how just about all the applications now are moving over to a cloud environment.
Then I'll turn the floor over to my friend and colleague, Dr. Robert Pepper, Vice President of International Government Affairs, who'll be joined by a distinguished panelist of experts, and from there, be able to talk about the regulatory impact of this growth that we're seeing, as well as the implications on the service providers business itself.
But first, to set the context, I'd like to talk a little bit about the Visual Networking Index as a whole. It started about 6 years ago, originally as an internal study meant to help guide our own engineers as to the levels of innovation that were going to be required to help keep our customers ahead of this rising bandwidth demand.
And the way we actually went about creating the forecast was to deal with third-party industry analysts, who had subscription growth forecasts. In essence, the number of people that would be subscribing to the mobile Internet, for example, in Germany or the number of the machine-to-machine communications that would take place in Japan. We then factored in, assuming that those subscription growth forecasts were true, what would the resulting impact be on the network as a whole. And then over the years, we've been able to refine this study, adding more and more unique sets of data from anonymous data feeds that we get from service providers around the globe looking at trends that are taking place within their broadband lines, as well as we have a whole panel of service providers where we're able to discuss with them the impacts of changes in the pricing models and the resulting effects that they're seeing. We factor those elements into our model, in addition to having a set of more than 900,000 users that are using applications that not only give them data about their own unique fingerprint or footprint on the mobile Internet, but also shares with us that data so that we're able to refine the forecast overall.
Now traditionally, whenever there is a question on our methodology, we err towards the conservative side. And we found that we, in general, tend to be anywhere from 2% to 10% conservative. So while some of the numbers that we present today may actually seem quite high and a little hard to believe, if history is the guide, they actually will be underestimating the scope of the true growth that will be taking place on the mobile Internet.
So with that as a background, let's dive right into the numbers as our Valentine's Day present to you and cover the findings of the 2011 to 2016 Cisco Visual Networking Index Mobile Forecast.
In the year 2016, we're anticipating that the total global mobile data traffic will reach 130 exabytes. Now exabytes are a large number in and of themselves. It's about 10 to the 18th power of bytes. 130 exabytes is an extraordinary amount of data. It's the equivalent of streaming about 33 billion DVDs going in the airwaves above us, so a massive amount of data. And it's also growing then at a very strong rate. Breaking this down on a month-to-month basis, we're moving from a level of about 0.6 exabytes per month over 2.1, which is 10.8 exabytes a month, which reflects about an 18-fold increase during this 5-year focus period.
Now for those of you who track the Visual Networking Index and follow it, and I know there are a number of you here as well as watching globally, you will see that these numbers do vary a bit from what we have presented in the past.
In fact, when we assessed last year's numbers, we had a forecast for what 2011 traffic would be, we estimated that it would grow at about 131%. But when we were able to look at the actual amount of data that came through, we found that we were under it by about 2%. In fact, the actual amount of growth in 2011 on the mobile Internet was 133%. We have gone in and made modifications then to the model accordingly to reflect this true size of growth. And as a result, you're going to see that it's -- the increase is not just from 2011 but incrementally through the years as well all the way up to 2015 which was the last year or the final year in last year's study.
Now you're going to see we're getting to a stage between 2015 and 2016 where there's a sizable amount of growth that is taking place on the Internet. While we are having a consistent stair-step increase, that last year is a particularly daunting one there because we're getting to actually the law of large numbers where even an incremental increase from a very significant base will result in a very high amount of data. In fact, it's 3.9 exabytes of data. And when we start to look at that in comparison, 3.9 exabytes of data, I mean, just the incremental amount of traffic that would be added to the network between 2015 and 2016 is about 3x the amount of data that's on the mobile Internet this year. So while it's increasing at a stair-step rate, that last step is definitely a steep one there.
Now consumers are very much driving this trend. That's very much kept pace with our expectations from past forecast about nearly 77% of all the traffic is generated from a consumer. However, it's getting increasingly difficult to segment what traffic is consumer and what traffic is business, as our overall work and lifestyles continue to blend.
When we look at -- this is a trend that is not just taking place in one region or another, but rather it's very much a global trend. Our -- Asia Pac is, in 2016, is actually going to be the largest consumer of traffic, taking up over 40% of the traffic. And that makes sense because they have the highest populations, so there are actually more users that are getting onto the mobile Internet each and everyday.
The Middle East and Africa has the highest growth rate, approximately 104% over this 5-year period and again, in large part because they're working off of a smaller base. Being that they have the least amount of traffic, it's easier to have that higher growth rate based on the law of large numbers. But the main takeaway is that this is not an isolated trend but something that is very much a global phenomenon.
Now this marks the fourth year that global mobile data traffic has more than doubled. And that's despite the fact that we're going through some uncertain economic times. That there's been an increase in the amount of offloaded traffic that service providers are continuing to invest in. And then also with the implementing of tiered pricing where -- many in the industry were expecting some decreases in the amount of traffic. The reality is, traffic is still continuing to grow at a very, very healthy rate.
Now when we look at offload 2011, we estimate that approximately 11% of traffic was offloaded from the macro radio networks. So that would be the cellular towers that really comprise the mobile Internet, and instead put on to other means or fixed access of connections. That could be anything from a picocell that exists in the stadium and airport to a femtocell like the one I have in my home, so that you're still able to keep that connection. However, the traffic itself is running over a fixed line. We estimate last year, 11% were offloaded. And by 2016, that percentage is going to double to 22% despite the fact that the amount of traffic is going to be increasing 18-fold. And this is really reflective of the increased amount of investment that service providers are making to help keep up with this rising amount of growth.
Now the users themselves are very much consuming more and more traffic. In fact, this year -- or last year, rather, about 0.5% of the population of the mobile Internet was actually consuming more than a gigabyte. We refer to that as the "gigabyte club". However, that amount is going to dramatically increase by 2016 with more than 60% of those mobile users. That equates to about 3 billion people will be a part of that club. And there are a number of reasons for that.
It's not just the 1% as some people thought that those -- that the very elite group within a service provider that would be consuming all the traffic. In fact, we started working with our panel of service providers to assess the impact of the top 1% on to the networks as a whole. And -- about 2 years ago, they equated to about 52% of all the traffic was consumed by just 1% of the users. However, now it's only -- they've been able to cut that consumption level in less than half. About 24% of the overall traffic is being consumed by that top 1% of users. And that's really reflective of increased scrutiny by service providers, changes in their pricing models and the like.
But the interesting thing is the top 1% isn't necessarily using less traffic. In fact, they're still growing in traffic. Top 1% are going to be having a growth rate of about 37% year-over-year over our study period. However, it's the top 20% as a whole that's growing at a much faster rate, so it's no longer just the small isolated set of people that are causing so much the growth, but rather it's a mainstream swelling of consumption that is taking place. So the top 20% actually has a very significant amount of growth, naturally reflective of the mainstream user base as a whole.
Now in 2016, we're estimating that there are going to be a total of 10 billion devices on the network, but 8 billion of those will be personal and 2 billion machine-to-machine. And if you break that down on a per capita basis based on UN estimates, that gives us about 1.4 devices per person for every person on the globe itself. And then knowing this audience, especially those in the room, I think we've likely already exceeded that amount. And that is a trend that's going to be certainly adding to more growth. The more devices that come onto the network, of course, the more bandwidth that then will be consumed as well.
Now laptops are really the dominant factor today. If you think about what laptops do on to a network, they have a large screen, so they're able to consume more video, and that video needs to be at a higher resolution. They also have very strong compute power, which means you can have multiple applications running at the same time. That means each of those applications can be consuming different streams of IP traffic simultaneously, which allows them to consume more and more.
However, you're starting to see a bit of a device-substitution effect taking place over the next 5 years, where there's going to be an increased amount of smartphones coming onto the network, while those in London may be used to smartphones, there is a great number of other parts of the world where they are much more rare. As though smartphones come onto the network and as they become more advanced to where they can run multiple applications at the same time and have increased compute power, they actually will be consuming the majority of the traffic, taking up approximately 48% of the traffic in 2016. And really supplanting laptops as the primary consumer of overall IP traffic.
Tablets are interesting story. They weren't even on our forecast 2 years ago. It hadn't really been launched in the mainstream. By 2016, we're anticipating that they're actually going to be driving about 10% of the amount of traffic. And again, that makes sense because they have higher computing power, and they also have a very large screen, which is optimized for video. In fact, if we look at the amount of annual traffic that will be generated by tablets in 2016, it's going to equate to more than twice the amount of mobile Internet traffic as a whole that was generated in 2011.
So just a single device that comes onto the market can actually have dramatic architectural implications for the service providers. And it makes me every time I look under a holiday gift tree to try to decide on what are the architectural implications, that 2 would be reflected on this one.
Now going over devices as a whole, you could see that they do range pretty dramatically. And as a result, any -- if you look at an individual service providers' growth rate or the amount of traffic they're going to be handling, it will invariably map into the type of device mix that they have on the network.
As I said, in 2011, those that have a great amount of laptops are going to have the highest amount of growth. Those that adopted the latest and greatest smartphones such as the iPhone, in particular, which consumes quite a bit of traffic, will have a very large base of traffic and -- that they'll be working from.
But as those devices continue to advance, that traffic is going to be moving over. So you can see from smartphones, they're going to be increasing, actually exceeding the amount of traffic consumed by -- in 2016 that laptops consume today in 2011. So this device substitution is something that providers are very much going to be working through.
Now it's not just all consumer devices that are going to be driving this traffic. Machine-to-machine is very much coming onto the mainstream as well. We're anticipating the traffic generated from machine-to-machine will be increasing 22-fold over this forecast period. But even more important than the amount of traffic that they generate, because some -- while some machine-to-machine applications such as video surveillance actually consume quite a bit of traffic, large majority of them actually are optimized to consume very little. They're tailored for that. In fact, it's much easier to plan for what machines are going to do on the Internet than it is for humans. And as a result, they're able to very much to keep most of the traffic impact for machine-to-machine at a manageable level.
But what's interesting when you're having an additional 2 million devices onto the network, it's not just the capacity you need for bandwidth, but it's also incurring a great amount of signaling challenges for the network, which is having to keep track of all these devices that are moving around. And also, because so much of the traffic has various quality of service ramifications, especially those that are dealing with video, it's incurring a great amount of need for not just capacity, but greater intelligence into the network as well.
Now as -- video, as I mentioned, is very hard to deliver. Our tolerance for bad voice is far higher than our tolerance for bad video. The network needs to know that this is a service that's going to cross the network and needs to make accommodations to make sure that, that experience that we have in watching video in the Internet is actually a good one. This has particular ramifications because by 2016, we're estimating about 70% of the traffic on the mobile Internet will be video-based. And this is very much in line with the trend that we have seen in fixed networks as well. Our network interactions are becoming increasingly visual, and video is the means by which that happens.
We're also seeing a rise in the types of applications on how they were using -- where increasingly they're becoming -- they're going more and more to the cloud. About 45% of all the traffic is generated from cloud applications in 2011. It's rising to 71% in 2016. And this is a significant impact on the overall traffic growth as well. Because when you download a movie onto a device, you effectively have a bandwidth hit to the network once. But when you use a cloud application such as in United States, a Netflix, or here in the U.K., a LOVEFiLM, every time you watch a movie, you're actually incurring that data stream. It's making an impact onto the network as a whole. And this is the same as it applies to music. We're not just downloading MP3, but perhaps, say, listening to Pandora as you're moving from -- even on your e-mail going to a cloud environment, everything is adding up to the cloud, and that's requiring constant network connectivity, connectivity that invariably eats up more and more bandwidth.
Now adding up to all this as well, we're having mobile network speeds increasing very dramatically. We're anticipating this 5-year forecast period that mobile network speeds are going to be increasing ninefold, which is quite a sizable amount of increase and really reflective of a great amount of the mobile CapEx investments that service providers are making around the globe.
Now from the growth rates in North America or Western Europe where they already have very fast mobile Internet broadband speeds, it's not going to be as high as that in, say, Middle East and Africa which again is starting off of a much smaller base. That's going to growing at about 97% year-over-year.
However, the overall impact of the traffic in North America and Western Europe is going to be pretty dramatic, having speeds in excess, on average, of 5.5 to 6.7 megabytes per second. And that's going to be increasing even further when we start looking into the rise of 4G when it comes onto the network.
By 2016, we're anticipating that 4G will account for 36% of the amount of traffic on the mobile Internet, makes sense because it's bringing about very high broadband speeds. And the faster you allow to people access things on Internet, the faster they actually do it. They are able to run more. It's going to be inspiring more of a mobile substitution away from your fixed broadband. And more and more, everything's going to be coming over to the mobile broadband.
Now the interesting thing about this statistic is that while 36% of the traffic is actually going to be from 4G, only about 6% of the devices are. So just as we saw laptops driving a great amount of traffic early on into the development of the mobile Internet, we're now seeing that these 4G devices that are coming onto the scene that will actually be some of the -- amongst the largest consumers of bandwidth as a whole, making for some very interesting times, of course, in the landscape of the mobile Internet.
In fact, each connection -- we were just looking at last year when we're in 2011, each of those 4G connections is already consuming more than 28x that amount of traffic of similar devices that are working off of a non-4G connection. So definitely promises to be an interesting time in the mobile Internet, and one that we will be most certainly to monitor in the next iteration of the study as well.
So to summarize, the overall amount of traffic we're anticipating in 2016 will be reaching 130 exabytes, 18-fold increase. Traffic is going to be offloaded at twice the percentage it is today, reaching 22%. There's going to be a rise in mobile traffic and of cloud applications that very much are going to be much adding to that. And then with faster speeds taking place, then it's setting up boundless opportunity for new ways to innovate and leverage the mobile Internet. It's opportunity that is both for the service providers to capitalize on, it also going to be a challenge for them to accommodate.
And so with that, I'd like to thank you for your time and ask my colleague and friend, Dr. Robert Pepper, to introduce the panelists to start diving into some of the changes then that we'll be seeing over the next 5 years from a business and regulatory front. Pepper?
Thank you very much, Doug. Fascinating data. One of the things that also in addition to the implications of cloud that we found in our cloud index that we released in November is the importance of low latency for cloud services to work. So again, this is another dimension of the implications for networks.
We're very lucky to have a very distinguished panel with us this morning -- or I guess maybe it's already evening in Asia. Hello, world -- to talk about the implications of the most recent data. We'll start with Gabriel Solomon, who is the Director of Global Affairs for GSMA, the GSM Association. And Gabriel has extensive experience in the international arena as well and is, in fact, leading some of the questions and programs within GSMA on digital device. So we'll hear from you about some of the implications globally, but also, for the operators.
After Gabriel, actually, we'll go to Martin Cave. Martin is actually here in his private capacity. Many of you know him as not only a distinguished professor and also a part-time government senior official. But Martin also is the author of the Cave report, which some of us actually remember as a watershed for spectrum policy here in the U.K. that then became a model for spectrum policy globally. And Martin, in fact, has just returned from Malaysia. So it would be interesting to hear what's happening globally. He also is a -- the part-time deputy chair of the Competition Commission here in the U.K., but he's here representing himself and not the Competition Commission.
Following Martin, David Hendon, who is clearly one of the U.K.'s experts on telecommunications and spectrum policy for many, many years. First it was DTI, then it was Biz. And David also led the U.K. delegation that -- at the WSIS that led to the creation of the Internet governance forum. David is one of the global experts on both fixed and wireless networks spectrum policy, telecom policy. And now in his part-time consultancy at Ofcom, he's focusing on international issues. But again, David is here in his personal capacity.
And then finally, Brian Williamson from Plum Consulting is going to talk about -- or has actually written recently a very interesting paper entitled, "Mobile Data Growth -- Too Much of a Good Thing?" And which is a fascinating question about business models for the operators. And Brian is going to start and talk about that. And then, we're are going to open up for questions, and I'm looking forward to hearing from the panel on what they see as the implications of the significant and increasing data growth across the mobile Internet.
So Gabriel, why don't we start with you?
Thank you, Robert. And may I also just take this opportunity to thank Cisco for your continued investment in the VNI, I think for a lot of people who watch the industries, as the de facto standard for forecast. And so thank you for your ongoing commitment. I think the numbers that you've shown today show that many -- for many operators, one of the key drivers to get mobile data usage into the market, the all-you-can-eat tariff is going to become an anachronism. I don't think it's a sustainable pricing strategy. And I think we'll see more and more operators going to tier pricing or value-based pricing, particularly as data shoots up or data usage shoots up in their respective markets.
We saw that you're predicting the "gigabyte club" going from 0.5% to 60% in 2016. Clearly, it's unsustainable if people consuming huge amounts of data are paying the same than people who are consuming few amounts of data. I think the second and perhaps ironic development that we'll see is that some of the proponents of the net neutrality debate are going to, I think, approach network operators and try and negotiate paid interconnect terms so that they can establish high quality of service, particularly for video, which is going to be 70% of the traffic across those networks because they will have an incentive for users for consumers video to experience, as was said by Doug, the video in a very low latency way. Otherwise, they'll simply turn it off. So OTT players will have an incentive, I think, to arrange commercial terms with network operators in the next few years.
I think also incumbent then on governments now is the release of new spectrum into the market, both the digital dividend and the 2.6 gigahertz extension bands. I think this is an increasing trend. We're waiting for it in the U.K., but I believe it's happening imminently. And I think it's important to try and align the release of both the digital dividend and the 2.6 gigahertz together because I think this then provides investors with a clearer picture of how to a, bid for the spectrum and value spectrum, but also construct business models to invest over the long term and to provide ubiquitous connectivity in a shorter period of time than perhaps otherwise possible.
I think also this massive growth, which you keep increasing your forecast each year, also points to the fact that we need to identify future spectrum bands. And this is a task that the ITE will carry out over the next period. But frankly, there will be no new harmonized spectrum coming into the market until 2020 at the very earliest. How cool was that?
Great, thank you. Martin?
Yes, thank you. I've been disseminating the results of this kind of work for a couple of years. And I must say, I find it very useful. In an attempt to try and galvanize policymakers, you might want to talk about very briefly and to do something about it. And it has been quite successful. The previous math were rather easy, I would say, is going to double every year. So after 5 years, it's going to be 32x what it is today. We've now learned it's not going to quite double every year, but it's still pretty terrifying.
And I think it's beginning to take effect in countries, which I have worked in recently, and which are interesting, but in some cases, quite small, Georgia and Thailand. And there's a lot of interest in trying to preemptively to put spectrum out in order to meet these needs, on the simple hypothesis, that it's going to be good for economic growth. And in Europe, of course, we've seen the Radio Spectrum Policy Programme. I guess I wouldn't have envisaged this happening. And when I heard it was going to be part of the package for telecoms reforms, I was reminded of an observation that Dr. Johnson when he compared something to a dog walking on his hind legs, it's not done well, he said. The surprising thing is that it's done at all. I won't tell you what it was he compared to a dog walking on his hind legs because it's 21st century, it's not a particularly [indiscernible]. But it was done well. I mean, that's the surprising thing. I think it's a remarkably good and ambitious document. I mean -- and for that to emerge from the European institutions with their rather haphazard methods of working is an indication of how these things have really taken hold. I mean in the U.K., we're planning to reallocate 500 megahertz of spectrum for the public sector alone. And all these plans are right -- are available with similar advantages, 450 in France, 800 somewhere else. And it's really taking hold. So I think that, that response has been quite encouraging.
One of the things that has made this possible to sort of close the loop, I think, is that in legislatures in the past, any battle for spectrum between broadcasters and telecom operators would always be won by the broadcasters. In the United States, they were the gatekeepers for Congress representatives to get on the televisions to speak to their constituents. They had enormous power. In the European Union, for slightly different reasons, the parliament was very pro-broadcaster for reasons to do public service broadcasting and that cultural, too -- cultural issues.
But all that seems to have been swept away in a quite remarkable way. It's a -- the loop is closed because all these new devices are taking people away from broadcasting as that particular means of access to political agents voters has rather come to an end, and I think that's quite surprising.
Now from a spectrum policy point of view, for proponents of markets and spectrum, as I have been for some time, one of the interesting things is that very little of this redirection is based upon markets and prices. It really depends upon the brute power of the government department or the regulatory agency. And such is the enormity of the shift that has to be accomplished in order to meet these needs, that, that's probably inevitable. It's like trying to decolonize your electricity supply industry. If you want to do it in 100 years, you can use prices. If you want to do it in 20 years, you've got to use sort of central planning methods. So it's a bit disappointing for me that we aren't seeing more of market methods. I guess in the U.S., they are at a greater extent with the incentive options and things of that kind. But it's working, and I'm very gratified by the way this policy change has taken place, and I think it's partly due to the scary data that Cisco has provided.
Thank you, Martin. And David?
Yes, thank you very much. I think it's absolutely fascinating actually to see this data, and I really thank Cisco for providing it. I sort of look back over the last few years, when I worked in governments and I'm now working at Ofcom, although as Robert said, I'm here on my own account today. But looking back, I can remember when we were doing the Digital Britain Report, I was very involved in that with Stephen Carter, and there was no suggestion then about the dominance of smartphones and the need to think about huge increase in mobile data capacity. But also no thought really then, particularly about the fact that mobile data was going to be so important compared with the fixed network data. And we -- I'm sure all of you are aware that like many administrations around the world, the British administration decided to go for a 2-megabit universal service, which I remember was twice as much as the Fins, which seemed like a good enough reason to do it all on its own.
And actually, now that just seems extraordinarily irrelevant. And one thing that immediately strikes me is that these numbers are astounding, that actually, in a couple of years time, we'll be sitting in a room as nice as this somewhere in the world, I expect, looking at numbers which I fully expect to be even more astounding. And so, actually, for planners, it's really difficult to proceed, it seems to me, because if you start planning for the numbers we've heard today, there's certainly going to be wrong by the time you get anywhere near there. In fact and I think Martin can take some comfort in the fact that the only response that is going to be fast enough will be a market response because I think it's impossible for ministers and even for regulators to move fast enough to lead the market, so to speak. The markets got to work, and I think the government and the regulators have to try to stop emerging market failures and do something about it before they get to be a real problem. But that is very difficult to do, and it's extremely helpful to have this input here.
So just looking at these numbers, just another couple of things strike me. One is that there's a huge increase in the mobile cloud. Now the cloud is, it seems to me, a wonderful solution to lots of problems. But in itself, it gives rise to all manner of new regulatory issues. And if you -- just looking at the moment, for example, all the stuff about the privacy settings on Google and -- or every time you log into Google, they tell you've got to read this stuff and so on. I mean, this is just the tip of the whole range of things about which users will have to worry about, who owns the data? Under what circumstances can they get access to it? How will it be protected? Will it always be there when they need it? All of those issues, which at the moment are important, but probably not the things that they're going to be worrying about in Downing Street.
Actually, when you've got -- I think it's 71% of the data will be the mobile cloud in 2016, that's extraordinarily important part of what makes life work. And so I think there are going to be all sorts of new challenges there. And I think also, it's going to be necessary for government to -- for the government here, but obviously, the governments everywhere to think about what these massive shifts from fixed to mobile networks mean in terms of their policies to bring sort of -- if I can put it like this -- modern communications to the least well-developed parts of their countries. Now in our case, the U.K. is pretty well-developed country, but there's lots of places that have really slow broadband speeds at the moment. The government's got a big program to sort this out, which I was involved in before I left Biz, and I think it will make a big difference. But that in itself is probably not going to meet the demand that the individual consumers will have in the future. And I fully expect that the cry for faster fixed broadband, which fills the post bags of my colleagues at DCMS at the moment will shift into a cry for faster mobile broadband in rural areas in just 2 or 3 years time. So it's really interesting. And then one last sort of thought is that this all, in a way, came from nowhere. And I just wonder what the next thing that comes from nowhere is going to do to this and -- because I really don't believe this is the last -- big disruption to the way that telecoms is working.
Thank you, David. That's a particularly good point. That should lead to some discussion. What is it that we don't know? Next, Brian?
I'll say something about some other ways of thinking about what might happen to mobile traffic and the implications of that, and I'll center on 3. The first is that -- and just to characterize this Cisco forecast. I guess it's is a bottom-up device-based forecast, and it's a plausible story. So let's try some other ways.
One of those is thinking about the value of mobile broadband and the device ecosystem and applications. And I think the key point which follows on from what David said is it's barely begun and we tend to forget that. We think the stuff has been there for a long time. It hasn't -- I remember the first iPhone, and that was not long ago. Now one implication of that is that the ecosystem around that is only just getting started. So it will become more valuable over time and that in itself will drive demand for data consumption.
It also means that there are things that I wanted to do that I can't do because other people don't have a device with the capabilities. And that was brought home to me when I wanted to meet someone for a rugby game -- at a rugby game. I thought I would share my location with them, then I thought no, I can't. He doesn't have a sufficiently advanced phone. So as more people get these devices, they become more valuable to each of us. So if anything, that would tend to kick up the growth rate over time. So there's a lot of value there, a lot of economic value, and that suggests that high traffic growth is plausible.
The other way I think about it is to think about how much data is involved with the activities we do today, and that a lot of that is in the home. There's a U.S. estimate that roughly 1% of total consumption of data in the home is online. It's a very small proportion. A lot of it's broadcast. It's shrink-wrapped disks for games. If all of that were online, it would be 100-fold increase approximately. And that's about a 2-year old estimate, but not far off. What percentage of total Internet traffic is mobile? I think again it's probably kind of 1% , 2%. So you have a factor of 10,000 times if you think of the things we do now as potential demand.
The third one, which is the one I focused on in the paper that was mentioned is what happens when you look at the cost implications of this data growth in terms of network capacity and asked the question, "Would people be willing to pay that much?" Which is another way of thinking about it. And the conclusion I've come to by looking at that is -- and it's a difficult thing to know how to get these estimates right but, yes, it does look plausible. And indeed, traffic growth rates above those projected by Cisco look plausible.
And that raises then the last question in a sense, which is, what strategy will mobile network operators pursue here? Will they welcome that growth, price accordingly and go forward in the kind of high-growth world, or will they tend to view growth as a problem and price to impede it and therefore end up in that kind of equilibrium. So I think there is some choice -- strategic choices here that will impact on the outcome over the next 5 years.
I'll stop there.
Thank you. So let me just pick up on your last point, which is, are operators going to view this as a good thing or too much of a good thing, right? And the extent to which they believe that they can actually manage the growth, right? And the question maybe to ask Martin, right, from a broad perspective of thinking about markets and competition, do you think that an operator actually will be able to manage the growth in the face of other operators or incentives for other operators to say, "No, we're not going to manage. We are going to still stay with the all-you-can-eat." For example, in the U.S., while we do have some of the mobile operators beginning to have tier pricing or spectrum caps, there are others that are still, in fact, are marketing their all-you-can-eat packages as a competitive response in trying to get people to churn and switch operators. So there's this competitive dimension which feeds into, Brian, your point, which is they're going to try to manage this. But do you think that's sustainable from what you've seen?
I can really only answer that on -- in sort of a very a priori way. I mean, obviously, the mobile sector is a small numbers game, there aren't [ph] hundreds of operators, limited amount of choice. So there's always -- I'll choose my words carefully here, there's always a tension between on one hand, a degree of tacit cooperation amongst the operators and on the other sort of out and out competition. And obviously, the more that any operator is actually running up against constraints on its Apple level because of, say, spectrum scarcity, the more likely it is you're going to end up with a more cooperative solution just simply because there's no point in competing and winning business if you can't actually satisfy that business. So it's very hard to say how that thing is going to work out. And I certainly wouldn't expect it to work out the same way necessarily in all the jurisdictions in which the scenario is going to be enacted.
Gabriel, do you have a thought on that?
I think there's -- I think operators see the growth as a challenge. Now it's a challenge to, frankly, make money through investment to provide services. Taking the point on U.S. people or U.S. operators advertising all-you-can-eat as a drill [ph], that's fine. But it's probably a short-term thing. I think fundamentally, it's not sustainable. I might be proven wrong, and that's good for consumers. But I mean to take -- this will extend Martin's point, I think despite perhaps in many markets there may be not being double digits amount of operators, there are nevertheless quite competitive forces at play. And I think all shareholders fund [ph] As investors want a sustainable model, frankly. And that's the route forward. And I think it will change. I think it will evolve. But I do firmly believe that the market is there, as was alluded to earlier, and it will find out the right solution.
So picking up on something that you said, David, which is this governments everywhere have a public policy imperative to make sure that all of their citizens benefit from connectivity. It used to be telephones, and now it's broadband. And with Digital Britain setting out some targets. And in U.S., it's -- we have our universal service fund and incorporating broadband. Do you think that there's an opportunity though with what we're seeing with the mobile broadband essentially leapfrogging the fixed broadband to not only make it possible, but it becomes standard that once you're in the air, right, you don't have to worry about long loops, high cost to farms and low-density areas. You actually have wireless networks that can reach areas. Does this make the problem worse or easier to solve?
I don't know that it changes it. It makes it more difficult to solve because the problem doesn't stay put long enough to really get your mind around it. And I think -- I mean, I think a lot of the final connections will be mobile always because why would you ever want to be tied to one place by bit of wire? It seems to me that telephony really took off when we invented the mobile sort of thing. And you can see the way the Internet has taken off just as soon as people worked to how to get the Internet into something that you could carry around. You didn't have to stay in one place connected to the broadband or whatever. So I think it will -- the end connection will be mobile. But I think myself that one of the numbers that I'm a bit suspicious of in this report is the mobile traffic offload, which is going to be apparently 22% in 2016. And I think that's low. Another number I challenged to when we were we were talking before was around tablets. And I said I thought tablets would be higher than 10%, but then someone said to me -- I forget who I was talking to at that time, said, "Well, though with tablets, you don't use the public mobile networks, you use Wi-Fi mostly, I think [indiscernible], actually." And that struck me as a really good point. And I've got an iPad, and I offload it as it were all the time pretty well, even though it's got 3G, and I hardly ever use that. So I suspect that there'll be more offload, and that's part of how we'll handle this data. It does strike me that this is an extremely disruptive time for the mobile operators. They've got to work out where the -- what the value chain looks like in the future. They've got to cope not just with this huge increase in data -- mobile data and the need to invest massively in networks to provide it, but they've also got to cope with the fact that customers are no longer just people that pick a phone up and use it. But actually, their customers are sort of hide behind apps. And actually, the apps are what connects customers to the value chain. And actually, if you're a mobile -- a network operator, I'm not quite sure how you can monetize the fact that people can put apps on your phone and pull data down. So I think all of this is very tricky, and well, I think probably actually it's more tricky even in the public policy response, which is in itself, strict enough.
Gabriel, did you want to comment -- or Brian on...
How will they monetize it? I think in the end, we will pay for connectivity and that will be the primary means by which my [indiscernible] monetize the applications will become independent to some extent of the network operator. But there's something in terms of what Martin said about competition, which I think is interesting. Because of the lack in my view of availability of tariff plans that are suitable for someone with multiple devices, and I certainly found that I got a tablet. I didn't want to take out a whole additional mobile contract. That's also encouraged people into this nomadic mode where they primarily rely on Wi-Fi for these devices. Now potentially, that creates opportunities for interim. So you don't offer a full mobile network, but offer something less than that. So I think the competitive dynamics are changing.
Picking up on that, just as a -- not to -- David, your point on the offload, it's an interesting definitional question. Between -- depending upon the country, the society, the culture between 65% and 85% of the use of a mobile device is indoors and sitting down and I'm not mobile. And if I want to think of mobility as high velocity and nomadic is a moving but at slow speed or walking around within a particular place, the notion of offload is the network operator and the device makes the decision to offload. When I connect my tablet -- when I actually do it, it's not considered offload. In fact, one of the things that we found in some other studies with VNI is that the majority of the way devices will be connected out to 2015 in this case is going to be by fixed wireless, Wi-Fi, and less so the fix Ethernet. And then there's the mobile Internet, which is also part of this. So when you look at the mobile Internet, 2016, 22% of that will be offloaded, which is really going to be the "network's decision" in addition to what you're doing, right? And your nomadic approach, Brian, where -- or David, where you're consciously saying, "I'm not going to turn on my 3G antenna. I'm going to connect it to my home Wi-Fi." You add those together, and wireless, both licensed wireless and fixed wireless, Wi-Fi in combination is going to be the vast majority of the way that people connect their devices to the broadband net. And that has implications per spectrum that's both licensed through the big macro cells, and then the unlicensed Wi-Fi or some of the license things like femto and some of the other micro cells. So we begin to put these together...
I think it also has an implementation for how we will think in the future about fiber because the fiber doesn't go to the device in the majority of cases. And so when we talk about these very high speeds, they're not necessarily the speeds you'll get at the device unless something else is done there as well in terms of the wireless environment.
Yes, and which you've heard me joke that people talk about mobile networks, I have no idea what that means. The network is not mobile, I am. The network is fixed. And if we're going to have these kinds of multi-megabit average speeds for the device through a radio connection or wireless connection, that is going to have to be connected to fiber at some point as quickly as possible. What I'm -- I'm sorry, David please?
So let's just speak -- you asked me about fiber and I think I was that quick. [ph] I think actually there's a huge future for fiber because I think what we need is -- we're going to need to have a low cost of ownership, reliable and scalable to cope with whatever comes in the future means you're getting traffic back from all over the place back into the networks, and at the moment, that's fiber. So I think actually the fiber rollout program, which we're doing in the U.K. and which other countries are doing is still hugely important. It's just that the way that people will interact with that or connect to which is going to change.
Why don't we open it up for some questions? We have one here. And then back there, there's a microphone with -- 3 so far. And if you can identify yourself? 4.
[indiscernible] Well, first question is a point of clarification. Are you saying that in the forecast, that you do not account for Wi-Fi offload as part of your mobile traffic growth.
No. The Wi-Fi offload, in fact, we have our expert who can, Chandra [ph], who can actually answer technically the question, but the
Yes. My understanding was -- Chandra [ph] why don't you just go ahead?
So our fixed Wi-Fi offload is included in our VNI fixed forecast. Today, we are talking about the mobile data forecast. So that is released in the June timeframe. So today, the mobile offload that we're talking about is primarily from smartphones and tablets. And that is...
The femto, pico, [indiscernible]
On the femto, pico and Small-cell, actually is Wi-Fi. That is correct.
Just a clarification, if -- can I ask you a question now?
To what extent in your analysis, do you actually build in things like natural constraints, i.e. the operator's ability to have the money to build the networks, the capacity of the networks to deliver and how much -- another way of looking at it would be how much of this is actually a wish list, if you like, of an unconstrained growth. But this is particularly important when we get to the large numbers that you are talking about, at the end surely that's impossible to sustain.
When we do our forecast, we are starting with the third-party subscription growth forecast, so these are industry analysts that assume the number of, say, mobile broadband subscribers in a particular region. And we put together this mosaic to get, effectively, a global picture. So those inputs that you're talking about actually are oftentimes factored into those. So we're just -- as those inputs change, so too will our forecast. We have actually seen 2 years ago, on our fixed forecast for example on -- via some economic downturn, there was a bit of a decrease in terms of the overall adoption rates of some services where we did see that decrease in capacity. This year, we're actually seeing a bit more of an increase.
I was thinking more of 4 or 5 years down the line. We've got this huge numbers. How much is it going to cost to build the networks to support that capacity?
We do not do a separate capacity study on there. Those are often times done by third-party analyst firms as well, so...
In the back here.
Hi, Dan. I'm [indiscernible] from Oden [ph]. I think it's a great piece of work, so well done, Doug and the team. I think we can argue whether the numbers will be 110, 130, 150 but the reality is it's going to be a huge growth. My question is for the panel, what does this mean for the industry? I mean, what is the value from this scenario, if you assume that these numbers will come true? Spectrum value increases substantially, so great for the government. We're going to sell a lot of kit, right? For vendors like Cisco. But what does this mean for the operators? And bandwidth becomes, eventually, a commodity, it already is. So where is the value in this ecosystem and how are the operators positioned in this environment?
Now, a great question. I think A.T. Kearny did a study a couple years ago where they looked at the economics of the Internet, and they found that 70% to 80% of the value was upstream, so with Google, with Facebook or whoever. And I think for this whole Internet ecosystem to be sustainable, I think some of that value have to flow down. And Vodafone followed that up with a paper on emphasizing on two-sided markets, the importance of two-sided markets. I think that's an area that we haven't seen evolve yet, but I think it's a very important one, and perhaps it's not going to be a huge amount of revenue coming in, but it will stem a bit of a tide of all the value flowing upstream. So that's one element that, I think, needs to evolve over the next few years. The other is this tiered pricing and being flexible with pricing maybe treating multiple SIM card owners differently, bundling things together, being a bit more intelligent about the way you interact with customers and extract value from customers.
And Brian, you talked about 2 different models and you think that both were sustainable.
Yes. I mean, we've done some of our own estimates, on our Ericsson heavy estimates on the public domain about the incremental cost of data growth. And it's not obvious from those that it isn't sustainable to see these sorts of levels of traffic growth. I mean that's partly because future network technologies are more efficient and cost less to carry each gigabyte. And it's also because we are making additional spectrum available, as Martin said, and we may make it more available in the future. And it's also because whilst networks are congested in some locations, overall they're underutilized, so over time, the total utilization will improve, which lowers the unit cost, again, of data. So it isn't obvious that it's a problem, and in general, for a business demand growth is a good thing. I can't see why demand isn't a good thing here. On the point about spectrum costs, again, it's not obvious. I mean, if we do implement the program that the European Commission are expected to pass, then not only the spectrum that were release in the near term but more in the future may be released. So it's not obvious that the value of spectrum at the margin is going to be driven up, supply side will respond as well.
Yes. Simon Saunders from Real Wireless and chairing the Femto Forum. The Small-cells were mentioned earlier on, and I think I have a semi-obvious role in offloading and providing capacity in places where it's needed. But I wondered, listening to the panel if there isn't a more subtle role for them? David mentioned the challenge in responding fast enough, and I wonder if the very point isn't the -- that it is uncertain, and operators need to be fleet of foot the foot to meet the demand as it pops up in different places and in different times. And in that context, I wonder if actually the dominant role of femtocells, microcells, pico cells and Wi-Fi isn't being able to respond very fast without the long cycles that it takes usually to roll out mobile networks from traditional macrocells.
Yes, I would like to respond to that because I agree with that. And I think actually the great thing about femtocells is that it enables the investment decision to be made by the person who wants the capacity. And they just go buy it and plug it in and way it goes. The planning department at the mobile operator doesn't necessarily have to even know this is going to happen. And I think the fleet of footness [ph] As it were is actually going to be an important part of how this is all going to work.
Two other points on that, one is that the uncertainty about the future is one reason on operator might buy spectrum and not necessarily use it. And I think we have to think about whether that is an okay thing for someone to buy some spectrum and not use it. Is that inefficient or is that the way they deal with uncertainty because if demand does take off, at least they do have that capacity. The other thing is with smarter tariff structures and -- for consumers. It maybe that I can change my behavior. If there's a shortage of capacity on the macro network, then you can steer me towards the femtocell or the Wi-Fi.
Allow me to give a point about Claren [ph] spectrum, not using it. The -- one of the things that we've observed in the U.S. is that operators tend not to do that because the opportunity cost of not using the spectrum is so high. And there've been concerns and the voice in the past when we started the path of spectrums back when -- I was at spectrum auctions, Mark, do you remember? About warehousing, and we've just really not seen that occur, because of this. And in fact, there is the -- especially where you have the ability to trade spectrum. And where you can have spectrum trading constraints, that's a different -- it's a different story. Question back over here and then here. Yes, go ahead. Use the mic.
Dean Bubley from Disruptive Analysis, very deep because I've got a cold. I have to say I'm convinced by some of the things I heard today, but I'll address those later on. But specifically, I'm interested in the panel's view on the role of changing user behavior because of pricing and because of device innovation. And also linked to that, the changing application behavior. So for example, on video, we see various content producers moving to things like adaptive bit-rate streaming where they adapt the content or the application to the constraints of the network rather than the other way around.
One of the things, by the way, that were -- with adaptive networks and devices and streaming is that there is the adaption to the form factor. So one of the things that we've seen is the data streams to the large form factor devices, actually, is greater than the streaming to my iPhone. So it cuts in multiple directions. Do you mind?
Yes, I think there are various channels through which user behavior or the bit-rate can adapt to fit with the network capacity. I mean the other point I'll make on that, I suppose, is that you have to think of both the technology changing and the behavior changing, because if one looks at the advances in compression, you might think they would reduce total traffic, but actually, with each advance in compression, you get a new type of application. For example, [indiscernible] gives you YouTube. And actually the change in behavior appears to more than offset the change in the compression technology. So it is dynamic, it is a complicated, but I come out of it still thinking that this kind of projection is plausible.
And when we talk about video being 70% of the consumption for mobile devices, and I think it's greater than that on a fixed network with the global fixed VNI. Video is not just, as we think of a traditional television, right? There is short-form video, as defined by Youtube of less than 7 minutes. There is video too -- there's long form video. There's video to the television, video to the mobile device. There is the 2-way video. So the, I guess, face-to-face or using our mobile devices with front and back cameras to talk to each other. So the notion of video also is very consumer-driven. So exactly your point. The application, its consumers are using their devices in new ways that were not necessarily anticipated, the disruptive point that you made, David. And also based upon the different applications tied to those devices. People are using Skype video over both fixed and now mobile devices. So it's very difficult to predict, and that's why the model is adaptive. And each year, we tweak it based upon the things that we've learned in the previous year. But your core point that it's consumers and applications, and the way they're using the devices is what's really what's going to be driving the demand.
It makes me think also in terms of a broader public policy debate. David, your point about what is it that we don't know? In 1984, Sir Donald Maitland issued for the ITU this global Maitland report on the -- on teledensity and the future of telecom and set out the global goal, right? Our goal is to have every person on the planet, no further than a one day's walk from a telephone, right? If somebody had said in 1984, 5.5 billion people will have telephones in their pockets, they would have laughed at you. And they would have said, "That's science fiction." And so just 1984, it's almost 30 years ago, but it's not that long ago that we've had this dramatic and radical shift in the way we think about connecting people. And the question is what's going to be next that we haven't thought about? I think we have time for one last question. We have 3 more though, so maybe Tony, why don't you...
Okay. Tony Lavender from Plum Consulting. Just coming back to the charts right at the start where you had the growth spike, 2015 to 2016. Clearly there's a spectrum policy implications depending upon what the trajectory is after that. And I wondered what's your feeling about -- is that a spike that then levels off or is that actually an acceleration of growth in day's traffic?
Well, what we're seeing is the again, the '15 to '16 is a large jump. In fact, the step growth that we're forecasting is actually larger than the previous years. At some point, it's going to level off. We haven't yet seen where it's going to level off. And again, these are global data. And as we -- and as more people are connected to smartphones globally, that is only going to continue to increase. In markets like North America and Europe, it may, at some point, start to slow down because of the penetration of the smartphone and other high consumption devices.
Yes. I'm just thinking at the level of uncertainty it actually leans on the end of your forecast, but...
Well, I'm not sure, but I know that Simon Saunders has done a study for Ofcom forecasting out spectrum needs for -- was it 2020?
Yes, 2030. And it was nevertheless -- it was not clear how we're going to meet those needs. Yes, we have 2 last quick questions, Nigel? And then, well, there and then...
Sure. John Delaney from IDC. Robert, you talked about how quickly things changed over 30 years, but one thing that hasn't changed is the amount of electromagnetic spectrum, and that won't change over the next 30 years either, that's finite. Given the massive increase in capacity that you're forecasting, and given the fact that I think it is Brian acknowledged that, essentially, operators can make money from connectivity to that network, is there room in any market for 4 separate mobile network operators in the long term or are we going to see massive consolidation at that level?
I think it's a general principle if you think that output is going to increase 20x. One wouldn't immediately jump from that to say therefore, there's going to be a massive amount of consolidation. I mean, whether it'd be entry, I think, is rather different and interesting question. But I can't quite see how the economics of the cost of mobile networks would change in the way that would force consolidation just simply because of growth.
I talked earlier about -- I thought the value chains would have to change for the mobile operators. And I -- really, what I was meaning there was that the business of building and operating a network seems to me to be no longer the essence of being a mobile operator, and we're already seeing that Vodafone and O2 are cooperating on building networks. T-Mobile and Orange, of course, have now merged, but they are cooperated very closely with Three on building network. So I think we could well see consolidation in network infrastructure. But that doesn't mean that there has to be consolidation at the sort of bundling, dealing with consumers and selling value point. And I suspect there might be a bit of consolidation, but I suspect there will be new players who'll come in new ways of making it happen, will arise as well.
Yes. And one of the things that I'm seeing globally is sharing a passive infrastructure. 80% of building out networks of course has nothing to do with radios. It's the backhaul, middle mile. It's building massive towers, security, backup generation for power and so on. And so one of the things that we're seeing is sharing of the passive infrastructure but maintaining separate ownership and competition on the active elements, so that you can actually reduce costs, but then actually differentiate your product in terms of the technology, differentiate the product in terms of different levels of service quality of service and so on. So that's, I think, affirming what David said.
I think another question is will governments allow consolidation? Vodafone tried to merge in Greece, it was not allowed.
And that is a competition authority question that we will not ask Martin. Nigel, last question.
I'll be very brief indeed. Nigel Hicks, Department of Culture, Media and Sport. First of all, an observation, in that a couple people have talked about the importance of this forecast, far be it for a government to decide whether these forecasts are important or not, but they are important in the context of how they're taken as evidence in the European Union. I mean, it was very influential, the Cisco reports, in terms of influencing the Parliament towards the position on the radio spectrum policy program that Martin spoke about. And it's also influenced in the ITU in the recent discussions in the last couple of weeks on the African proposals in Geneva to allocate more spectrum for mobile broadband. So I think there's policy implications there. But the -- I suppose the question was whether the forecasts, which are exciting but also worrying in the sense of how policymakers and legislators will react to them, take into account this. There's also a trend we're seeing certainly in Europe and also globally to an extent where the Internet is being disrupted or the potential of the Internet being disrupted in terms of data protection legislation, in terms of possibly net neutrality legislation, and how that could affect the forecast, I suppose is -- it will be an interesting question.
Anybody here wants to comment or -- so I do think that to your point, Nigel, that I appreciate the comment about the influence of the data. I think that it's extremely important that we be focused on continuing the fundamental open policies of the Internet and free flow of information. There are proposals, not just in the context that you've raised, but there are some proposals also at upcoming global meetings over the next year that could have very significant effects on the Internet as we know it. The Internet has been an extremely powerful, important and positive force in terms of economic development, spreading social benefits across the globe, closing the digital divide, making opportunities in education and other social purpose applications available. Proposals that would either balkanize the Internet that would break up the global nature of the Internet, I think could have a very deleterious effect on all those benefits. So I think we all need to be watching those things very, very carefully. And so I think you make an excellent point. That's not to say that we don't need to address important questions about making sure that the trust in using the Internet increases. And in order to increase the trust in the Internet, people have to feel secure that their data is secured. That in fact, they are not going to be subject to fraud or hacker attacks. So it's a balancing, so trust is important, but over -- in my opinion, overreaching regulation or some of the proposals to effectively balkanize and break up the end-to-end global Internet would have a very dramatic and negative effect. So thanks for raising that.
Doug, why don't we turn it back to you to close out?
Good. All right. Well thank you very much and I'd like to thank our panelists for providing their insights, and the ramifications of the forecast that we had today. I'd like to thank you all of you in the room for coming and sharing this global launch of this forecast with us. And of course, I'd like to thank everyone who is with us online. I'd encourage you to go to go to cisco.com/go/vni where the full details of the report will be there. We also have some nice interactive tools where you can get regional data. You can break things down by service or application or by device. We even have a number of countries there that you're able to pull out additional one. I encourage you to use it. I hope you find it useful for you as well.
I know there's some industry analysts that we have Elizabeth is waiting to see at the end of the day. But I would like to very much thank everyone again for being here. And please stay tuned for -- in June, when Cisco launches the fixed Visual Networking Index forecast. Until then, we look forward to hearing your questions and thank you again for coming.