Cisco Systems, Inc. (NASDAQ:CSCO) Wells Fargo Tech Summit (Transcript) December 4, 2018 4:30 PM ET
Ish Limkakeng - Senior Vice President, Product Management, Data Center Business Group
Aaron Rakers - Wells Fargo Securities, LLC
Thank you all for joining us after lunch. I appreciate it. I'm Aaron Rakers. I'm the hardware, semiconductor and com equipment analyst here at Wells Fargo.
We’re excited to have Cisco, Ish Limkakeng – I think I got that right – who is the Senior Vice President, Data Center Networking.
Before we start, though, I've been asked to read a quick Safe Harbor. Cisco may make forward-looking statements, which are subject to risks and uncertainties outlined in detail in our documents filed with the SEC, including our most recent Form 10-K, 10-Q filings. Actual results may differ from statements made today. No new financial information regarding Cisco's overall performance is intended or implied, and nothing in this regard should be viewed as an update to the current quarter.
With that said, I think I got that right. Why don't we kick it off maybe just to give everybody an idea of your role at Cisco, maybe a real quick background, and then we can jump into questions and kind of go from there.
Yeah, absolutely. Thank you. Thank you for having us. And thanks Wells Fargo. This is a great venue.
My name is Ish Limkakeng. I am Senior VP for our Data Center Networking Group. I have headed a variety of roles at Cisco, including working in our enterprise networking and the catalyst switching business. I have also been in the sales organization. And for the last, I guess, it’s almost 6 years now, I have been in the data center networking space.
So, it's been a very exciting time in the data center and, of course, the cloud. And I was part of the team that brought to market our Nexus 9K and our application-centric infrastructure, or ACI, portfolio.
And so, I now have responsibility for the entire Nexus portfolio as well as the other aspects, both the analytics and assurance platforms. That's what I do today.
Perfect. I'm sure you're not that busy. So, maybe we start the discussion by – I think any conversation I have today and probably over the next several quarters, if not longer, in the 2020s, is going to be around the positioning around 400G, how you see that playing out. And maybe just touch on Cisco’s portfolio and how, if at all, maybe the positioning around 400G might be a little bit different than that what we saw in the 100G cycle and how we should kind of go from there.
Yeah, absolutely. So, you probably have seen in the last 3, 4 weeks, maybe slightly longer, if I have the timing exactly right, we've announced our next generation data center, 400 gig products.
Some of our competitors have done the same. And I think it's indicative – it's a very exciting time in this market. If you look at every generation when we’ve made the next speed, there's always – when is the silicon and the switching piece of it ready and when is increasingly – as we get to the higher speeds, when is the optical component of it and the technology ready.
Those two things are kind of coming together now in 400 gig. And that's why you see the product announcements. And so, if I talk about each independently, we expect – we've got some early trials with customers you'll see. Of course, some of the big Web players be some of the early adopters, but there's also some service providers and even some pretty forward-looking enterprises that are starting to think about it at least.
So, we’ll have product here. I think we announced part of our announcement early next calendar year. We expect those to go to trials and then, of course, start to ramp towards the back half of the next calendar year.
The other piece of it is also the optical component of it. And I think what is – this is something, to your question, of what's different from the 100 gig cycle is that the industry has now more or less coalesced around a single standard, the QSFP-DD standard, which is good because I think it's something, which I'll talk about in a moment, that we were able to work with a lot of the big potential customers to understand their requirements and make sure that this standard was the right one for them, and it helps us to work closely with those customers.
But what makes it different for the industry is that, in 100 gig, we had many different formats. And so, it took a long time, frankly, for the optical component industry and, of course, the equipment vendors like ourselves to understand and drive volume and the cost to make it commercially viable for our customers.
So, I expect that will happen faster. Just go around because the industry has kind of coalesced a lot sooner. So, I think that's something that's kind of fundamentally different from an industry’s perspective.
And you mentioned QSFP-DD optics. There's also been some – one of your competitors has mentioned their support for both, QSFP-DD, but also OSFP. So, maybe help us understand of Cisco's views of why QSFP-DD is kind of the key driver of this industry adoption versus maybe OSFP.
Yeah. So, I think, again, working very closely with a lot of the potential early adopter customers enabled us to understand that the characteristics of QSFP-DD, without getting into a lot of physics, it has some very interesting characteristics from a backward compatibility and use of existing fiber plant that is not necessarily true of OSFP.
And while there are some advantages to OSFP, I think we were able to conclusively demonstrate the DD format really last – I think it was February or March in the optical conference. So, the potential concerns went away. And once can demonstrate it, the industry can move forward. So, I think that was kind of a key moment in that kind of coming together.
Okay. And so, when we think about the product cycle and the setup now with the products announced, is Cisco of the view that we start to see volumes in the mid part, second half of 2019 or do you think that – is Cisco’s view much more of a 400 cycle as a 2020…?
I think that's fair. I think, early on, we’re going to see a lot of trials and such in the first part and then we’ll start to see it becoming a little more material in the second half of the calendar year 2019.
Okay, fair enough. Maybe related, but maybe shifting gears a little bit is, I've seen articles where you’ve given interviews and you’ve talked about kind of this go-to-market evolution that Cisco has had around hyperscale public cloud, the importance, I think, of end markets growing at 15% or 20%. The traditional enterprise is much slower. And so, how has Cisco kind of changed over the last 12, 18 months in terms of inserting itself as garnering more and more footprint in those big, hyperscale vendors?
I think that's right. I think we have one of the things that we – we really kind of changed our approach a bit. And the realization is that all of these big hyperscale players are – in a very real way, they're trying to do things that have never been done before. They're building infrastructure and networks and control capabilities at a scale that just doesn't exist.
And so, what that realization drives is that our engagement with some of these players – and they're all kind of a case of one, so it's a little bit different with each one. But our engagement had to change a little bit from more of a traditional business development go-to-market, how do we engage at a business level and change it more to a direct engineering problem-solving model.
So, this optical format, for example, is a very good example of that engagement rather than kind of working either in a vacuum or working with some of just the optical component companies to think through the standard. We actually work very closely with a lot of potential early customers, co-chaired forums, did a lot of technical work together to establish that this would be something that met their needs.
And so, that kind of problem solving versus probably a more traditional go-to-market model, I think, is an important difference.
Is there anything that you can point to that kind of exhibits the success that you’ve seen in taking what you just laid out and actually seeing results from that in a data-thirsty world?
I think if you think of the big Web scale players, I think we’ve been public in the last year about some success we had at Alibaba. We've been public about some of the success in the work that we've been doing with Microsoft and SAI, their switch abstraction. We participate in Open Compute forum. It's a little bit of a different part of the relationship, but we have a partnership with Google on some of the work with Kubernetes.
What those things all have in common is that there is a very real technical problem-solving component to it versus a more traditional kind of business relationship. The other part of it is there too, of course, but that's kind of a different change in the axis, if you will.
Definitely, if there is questions in the audience, feel free to raise your hand and we can take those, as people kind of think about their questions, I'm hoping.
Maybe I’d like to shift a little bit. One of the things that we’ve seen in the Cisco story has been this kind of model transition that Chuck has kind of driven in terms of driving a subscription recurring revenue stream.
I'm just curious of how – maybe starting at a high level, how that’s played out in your business and probably even, more importantly, how I or investors should think about that playing out over the next 12-plus months?
So, I think – and I'm sure everybody is well aware, when Chuck took over as CEO, one of the key changes that he made relatively early on is put a stake in the ground that we’re going to become more of a software recurring business model.
And you shouldn't underestimate the idea of, like, hey, that's the goal, we’re going to go o it because that changed a lot of the focus inside of the company, from product development to go-to-market, I think you see some of the things that we’re doing in customer experience now and investments we’re making there is all related to that stake in the ground.
I think it's also helped by the fact that many of our customers have, over time, become accustomed to as-a service models, whether they consume classic enterprise today, may have 2030 SaaS providers that, obviously, is consumed in that fashion.
So, I think, overall, you have a lot of different businesses within Cisco that are in various stages of that. You’ve seen some of them, like the collaboration and the security businesses, probably more naturally fit in those models. But we have also started to offer that about a year ago in the catalyst business, subscription model, and we’re very happy with the success of that. And we've also, earlier this year, offered that across the data center portfolio. So, it's a little bit early in my particular portfolio, but we have a lot of great interest so far from our customer base.
Okay. That’s great. And maybe just to help me understand a little bit, when you say that you’ve implemented that across the data center portfolio, what exactly? Is that DNA? What exactly have you done differently and how I should think about that because you've been very successful in catalyst?
Right. So, think of it – the way to think about it is the same model, but we brought it to the data center portfolio. So, everything that we have from a software stack, we will also offer in a subscription model. And so, whether that is our NX-OS, our Nexus software, our ACI or our Tetration analytics platform, all of these are also offered in a subscription model.
Yep. And that kind of brings me to the other topic, ACI, you've been responsible. I think you came from Insieme.
Yeah. And so, I think the – maybe just a quick update on the ACI strategy, the traction that you've seen there, maybe other things around incremental points of integration across the portfolio that we should be thinking about looking forward.
Right. So, when we introduced ACI, the idea was that let's think about – it’s our entry into that software-defined networking market. And this is five years ago. We just actually had the five-year anniversary for ACI. So, we’re very proud of that.
The fundamental idea was use the techniques of SDN, but build it for what customers want, which is, at the end of the day, you only build infrastructure – put workloads in cloud because you have some application that runs your business that you care about and then there's data that that application uses.
It’s not about building beautiful networks. I've been in Cisco a long time. I’d love to tell you it's about building beautiful IP networks. That's not why customers build data centers. They have applications they care about.
So, it's a very simple idea. Let's build our SDN controller in a way where a customer can express – instead of using a programmable interface to create VLANs and VRFs and IP addressing, all of these things which are important networking constructs, instead give them an interface where they can say my application tier can talk to my Web tier which can talk to my database tier through these set of firewalls or load balancers or whatever constraints they might have. And it's a very simple idea, but it drives a very different development process and a very different outcome for the customer because they can express the intent of the application in one place and drive the infrastructure.
And we've taken that concept into what you see in the catalyst and intent-based networking, but it’s the idea that you tell infrastructure, at a high level, what you want it to do and the infrastructure kind of deals with it.
Q - Aaron Rakers
Yep. Any audience questions? Go ahead. Let’s use the mic real quick.
Thank you. Just back to the subscription services, I'm curious, at a business level, how are you incentivized to execute that sort of enterprise-level strategy at a business level?
Yeah. So, a couple of things. It’s clear direction from Chuck and Kelly, and you shouldn't underestimate that because we have regular review of our results and the questions come, like what are you doing to further this overall goal.
And I think – so from an offer standpoint, we look very hard at everything that we introduce and the existing portfolio. Does this make sense for – and do customers want to consume it in a subscription model?
And because of that dynamic that I had mentioned before, a lot of customers have kind of come to the idea that it makes more sense for their capital allocation and for their cash flows to kind of manage it that way. There is a significant demand. And so, that actually, by itself, drives a lot of the behavior that you might expect.
Is it 100% of customers? No. There's plenty of customers who continue to consume in a more traditional manner, but I think it’s something that we see consistently from a lot of customers for us to make those decisions.
Any other questions? When we think about the success of the catalyst side of the business has had in that go-to-market model, should I think about the data center side just taking a little bit longer to drive that adoption of a subscription purchasing model? Or maybe is that not the case, we’re just earlier innings that we’re at…?
I think we’re in earlier part because we did introduce the subscription piece specifically in catalyst ahead of where we are in data center, almost maybe not quite a year, but about almost a year in advance. So, we are earlier in that cycle.
I believe that there is a lot of appetite with our data center customers. I think it's a little bit TBD to see if it is the same level of appetite because oftentimes it's a different buying center between data center and a campus network buying center. So, we'll see if that plays out in the same fashion. But I think it's more, to your point, a little bit earlier in the cycle than it is fundamentally different, at least from what we could see right now.
That’s helpful. One of the things that we've seen kind of with one of your competitors is kind of this idea of kind of bridging from a data center to a campus strategy with kind of a singular operating system approach. I'm curious, as you've seen success with ACI, you kind of about what you've done software-defined, is that kind of an architectural evolution that you see that's kind of where your business and that of the adjacent campus business starts to kind of melt together a little bit more?
I think we see it a little bit differently. If you think about what's important in the campus, a campus network is all about connecting users and devices and access control for those users and devices.
And so, if you think about a heavy component of wireless being the most obvious example, right, which you don't have in the data center – data center is more about workloads and assets and the traffic patterns are a little bit different. It’s a little more north-south in the campus. More horizontal, east-west, machine to machine in a data center. So, we kind of have to look at these as – there are some different requirements.
Now, that being said, the idea of driving infrastructure in the data center with ACI from an application point of view and the idea of driving in the campus around this idea of user access and how you'd – software-defined user access is very similar or even the same concept. And so, it just has different aspects of the architecture and different – a little bit different focus.
So, if you look at what we do in the campus, there's a lot of wireless and power over Ethernet and voice and that kind of a focus versus an understanding of virtual machines and containers and services in the data center.
Okay. That’s helpful. We've talked a lot about, at the beginning of the discussion, around the 400 cycle, 400G cycle. But I think people often lose sight of where we might still be at in terms of the 100G cycle.
And so, I'm curious, as Cisco looks at their footprint, their install base in the data center, where do you think we’re at in terms of the upgrade? How much of your infrastructure is still 10G for that matter and not even yet had moved to the 100G cycle?
I think there’s still – of course, there’s still a lot of 10 gig server connectivity. I think you will see one of the – some of the early adoption of even the 400 gig might be to populate it with multiple 100 gig and breakout. But it's – that's actually one of the benefits of the standard that we’re behind, in that there’s some – there's a backward compatibility and customers have that flexibility. So, I think while I wouldn't call it necessarily early in 100 gig, there’s still a ways to go there as well.
Any other questions real quick from the audience?
I was wondering, could you maybe talk about your kind of strategy or views around merchant versus custom silicon on the switch side for data center?
Sure. Thank you for the question. So, we have had really a pretty consistent view on this for a very long time, which is we are – we use the best tool available. And, actually, that doesn't extend to just merchant versus custom silicon. That's true for whether we do things in silicon versus doing it in software. It’s kind of the best tool available. So, we’re one of the largest, if not the largest consumer of merchant silicon – at least networking silicon in the world. And we, of course, also develop our own silicon. It's not a religious thing either way. It's just what's the best tool available.
And so, the places where – and we continue to do this. The places where merchant silicon does a great job, doesn't make sense for us to spend our R&D dollars to replicate that. But in places where we believe that, based on what we know about the market – and we have, obviously, a long history in ability to develop on our own – if we can innovate and differentiate ourselves, some of the things that we do in ACI is a good example of that. Some of the things that we do with the Tetration analytics platform is an example of that. Then that's where we’ll invest our silicon resources and make that distinction. But it's kind of the best tool for the job.
Maybe I’ll kind of – in the two minutes or so we’ve got here remaining, maybe take it up a level and kind of think about what you might be seeing, if you could touch on it real quickly, in terms of the end market demand dynamics of your business, be it the public cloud guys versus maybe the traditional enterprise versus maybe the telecom service provider space and how you might be thinking about how 2019 sets up right now?
So, I think we – and this has been true for a number of years on the public cloud side. There continues to be pretty healthy growth and significant buildout. And I think that's probably not news to anybody in the room who’s following this. I don't see that changing.
There are, of course, a handful of very large providers. And so, at any given quarter, there might be a little more, a little less because there are relatively few providers making decisions about how they buildout, but, in general, that buildout continues.
I think one of the things that I believe has changed is the point of view on cloud and what it means for enterprise networks. And I think there’s probably general agreement that hybrid cloud and hybrid scenarios is really the foreseeable future. And so, as enterprises expand in the cloud, there is still growth, not as big as the public cloud, but there’s still growth in the enterprise segment for us.
And then, finally, I think you said the service provider piece. Actually, also pretty healthy growth segment. I think, of course, one of the big questions and one of the big trends in that market is the understanding of where 5G is going to go and what that means for our service provider customers, not just their networks, but their business models and how they're going to monetize that. So, I think that is one of the big questions in that segment.
Okay. I think we nailed it on the timing side. Appreciate you taking the time to talk to us today.
Great. Thank you very much.