Clash Of The SD-WAN Titans: VeloCloud Founder Uppal On Huge Customer Savings And Why Legacy Router Vendors Should Feel Threatened

Sanjay Uppal tells CRN the pace at which enterprises are deploying SD-WAN is unprecedented in the history of the market, and is rivaled only by the cost savings driving the rapid adoption.

The Growth Game

As the founder of VeloCloud, the SD-WAN market's top player when it was acquired by virtualization kingpin VMware last year, Sanjay Uppal has a commanding view of a market poised for rapid growth.

For Uppal, the pace at which enterprises are deploying SD-WAN is unprecedented in the history of the market, and is rivaled only by the cost savings driving the rapid adoption. Customers see a 3X cost improvement when they move to SD-WAN, Uppal said, while liberating them from the ties of proprietary hardware and software.

One of the key pieces of hardware coming under threat is the router, Uppal argues, saying legacy router vendors are being forced to play a tenuous game to maintain those legacy businesses while simultaneously pushing the very solutions that could spell their demise.

"Router vendors and anyone selling complicated, custom hardware that goes on the premises of an enterprise has got to be concerned because the revenue stream they have right now is threatened," Uppal said. "Any router vendor would be looking at this saying, 'wait a minute, if I deploy SD-WAN, a lot of my custom hardware business is going to get threatened.' But they cannot afford to not do it, so they're caught in the usual innovator's dilemma, which is the more they push SD-WAN, the more they're going to impact their legacy business."

What follows is an edited excerpt of Uppal's conversation with CRN.

How much do customers stand to save when they make the transition to VeloCloud SD-WAN?

There's a simple answer, and a much more detailed answer to that. We've have about 2,000 enterprise customers right now. We've done an analysis, we've had a third party do an analysis from a TCO standpoint. On an average, you get a 3X improvement. There's a lot of variability, though. There are three parts to the improvement. You get savings from an opex standpoint. No longer do you need all the people to be able to configure CLIs. It's a dramatic change in the number of people you need. Second is a reduction in capex because you don't need custom, complicated hardware at each one of your locations. The third is the potential savings from the transport. You don't have to have only a private network. You can mix private and public and the difference between MPLS, broadband and LTE. Depending on geography, those things may be very different. In Europe the pricing differential between MPLS and broadband is not that high. In Japan, it's not that high. In Africa, it's huge, much higher than what it is in the U.S. The savings are different, but on average it's a 3X improvement. SD-WAN is taking off because there's a bottom line impact to the enterprise.

Are routers the legacy hardware that's most threatened by SD-WAN?

Yes. The first would be routers. The most prevalent hardware out there at the edge of an enterprise is routers. The first piece of hardware that gets taken out is not the router, interestingly enough. It's the WAN optimizer. When you put SD-WAN in, it is solving a problem that basically eliminates the necessity of having a WAN optimizer. The WAN optimizer, if you compare what is the cost, the WAN optimization is the most expensive, so it makes the most sense to take the WAN optimizer out. But it's not the most prevalent device out there. Then your VPN device would come out, potentially your firewall would remain there depending on how you're rearchitecting your security. The firewall physical device might come out. Then the router could come out as well. The reason the router comes out a little bit late is the WAN is the lifeblood and people want to make sure everything is perfect before they take the router out. We're totally fine with that. We can coexist with the router and over a period of time as the IT department becomes more comfortable with SD-WAN and VeloCloud, then we take the router out.

Does that make it more complicated for Cisco to get into SD-WAN than it is for VMware?

Yes. In general, router vendors look at this as both a defensive and offensive strategy. From an offensive point, router vendors and anyone selling complicated, custom hardware that goes on the premises of an enterprise has got to be concerned because the revenue stream they have right now is threatened. It's the same thing we saw in other areas as well. When Amazon moved a lot of compute to the cloud, complicated, high-end hardware on the premises is what got threatened. Anyone doing complicated, custom hardware for wide-area network services, those are the ones that are going to be threatened. Any router vendor would be looking at this saying, 'wait a minute, if I deploy SD-WAN, a lot of my custom hardware business is going to get threatened.' But they cannot afford to not do it, so they're caught in the usual innovator's dilemma, which is the more they push SD-WAN, the more they're going to impact their legacy business. As far as VMware is concerned, we have no such issue. There is no conflict there. I call it conflict-free go-to-market. There is no product within VMware that gets threatened when we put SD-WAN in. It's even a benefit. As we put SD-WAN in, the network is getting virtualized, and you can put more virtual services on top and you can sell more VMware.

To what extent is the competitive landscape heating up, and to what extent is it an API battle?

I have a deep respect for my competitors. There are a lot of smart people out there that have looked at SD-WAN and have come up with their own solutions. We believe ours is the best, but there are many players out there. The battle between virtual services run on the cloud vs. people trying to do on-prem better, an incremental improvement over what was done in the past, that's really what the battle is. There are folks that are existing router-based players who are always trying to protect their legacy, even though they might say they're channel-first. The size of the market is so large that it will attract investment and there are many players out there and I believe there will be consolidation as we move along… API battle… Is it a platform battle? Yes. Long-term, it's really the company that has the best platform vision that is going to win. SD-WAN will solve problems today, no doubt. How do you really take the fact that there's an overlay technology and as you go you can put on virtual services from security to analytics to machine learning to artificial intelligence… the company that comes up with the best platform is going to win. VMware is the leader in the virtualization space. That's what they've done on the compute side. We're repeating it on the network side.

How long will SD-WAN and traditional routers will coexist?

In the WAN space, technologies last for a long time. It's not going to be that routers disappear tomorrow or that MPLS revenue is going to go down. That makes for good stories, but there's going to be a fair amount of time for infrastructure to change out. We're talking about a fundamental shift taking a piece of hardware that's been locked on the premises of the enterprise for years. It's been upgraded, new releases have come in, people have been trained. Now, we're saying forget that, move it into a service that is running virtually from the cloud. You're changing something that is a physical product into a virtual service. That's a two-pronged change. Along with that is all the change that is going on within the companies themselves. People change, training changes, and usually that takes a period of time. The move inexorable. It's not going at any slower than any of the other technologies moving to the cloud, whether it's compute moving to the cloud, or applications moving to the cloud or storage moving to the cloud, networking is happening at at least the same pace as those others have happened.

Is the revenue opportunity for vendors and channel partners just as big as it was for those other movements to the cloud?

Yes. Actually when you look at it, because this is industry-changing, depending on how you count it, it's a $45-$60 billion market and what we're seeing is that MSPs in particular, and VARs that are becoming MSPs, are the ones who are going to benefit from this, and they're already seeing that. Included in MSPs, I would put service providers that can be both a sell-to and a sell-through for SD-WAN. These are the folks who are going to benefit. The enterprise does want to outsource. They want to buy things on a pay-as-you-go basis. They want a subscription. We've seen that with the other three phases, as well. The really large banks resisted the cloud for a long period of time, but even they are coming around and saying yes, we want some portion of our IT in the cloud. Networking just becomes a part of that, and the people who benefit are the channel partners like MSPs. VARs that only provide boxes are going to be threatened, but we're finding that almost all the VARs we touch are either moving to become MSPs or have MSP arms, or MSP groups within them. I'm not seeing anyone who says I'm selling boxes and I will only continue to sell boxes forever. VeloCloud is 100 percent channel, and we'll continue to do that at VMware, as well.

Has there been a bigger revenue opportunity for the channel in the last decade?

MSPs have benefited. Service providers have lost ground, but have gained ground in other areas. Overall, everyone wants to move up the stack, to touch the application layer, not just the networking people, but all the way into the strategy offices, the CMO, the CEO of these enterprises. At a large retailer I visited two years ago. I walked into the meeting and they were talking about several thousand branch offices, and they were presenting this new strategy of SD-WAN, and then the CEO walked into the meeting. Later on, they said the CEO has never walked into a meeting where wide-area networking was being discussed. It illustrated to me the shift that is happening and how it is becoming relevant to the strategic part of the house for the enterprise. Because the product is becoming a virtual service, anyone who aligns themselves with the virtual service will see the benefit. Several partners have jumped on it from the early days, and they're reaping the benefit. You're going to see it touching other parts of the partner network, as well. VeloCloud has become part of VMware. Dell is very much part of that equation. From a distribution standpoint, when this becomes really widely available, then two-tiered distribution comes in and that part of the partner equation will also start benefiting from SD-WAN coming in.

How important is the SD-WAN market today? Is it on the same level as virtualization 20 years ago?

From in industry standpoint, in my view, WAN is a piece that cannot change very often. It can't change very often because it's the lifeblood of all businesses and you don't change your lifeblood very often because there's a lot of pain when the change happens. The fact that SD-WAN is gaining so much visibility and this many deployments this early in the lifespan is because there are very dramatic benefits to putting SD-WAN in. You look around and see the pace at which enterprises are deploying SD-WAN and it's approaching 8,000-9,000 enterprises now, and that is unheralded in the WAN space. It is the most important networking change to happen in the last few years.

How are SD-WAN and software-defined networking related and/or distinct?

You can think of it as part of SDN, because there is a part of it that is about the separation of the data plane from the management and control plane. Other than that, from an architecture perspective, the business model, the business case, the reasons for deploying SD-WAN are very different from deploying SDN. That's why the pace at which it is being deployed is very fast. SD-WAN is absolutely as important as virtualization was years ago, or any one of the other trends going on, like IoT, and they're all related. You cannot distinguish SD-WAN that clearly. There's an overlap with IoT. There's an overlap with virtualization, with machine learning and AI.

And it does that while liberating customers from proprietary hardware and software?

We look at it as not just a solution for problems that enterprises see today, problems of complexity, cost, application reachability and performance. We also look at SD-WAN as a platform, very similar to virtualization. When you had fixed hardware and software, virtualization came in and provided a new layer that freed you from all the things that could run on top of the virtualization layer. That's very similar to SD-WAN. SD-WAN comes in and really frees you from the specifics of the transport layer so you can run all kinds of virtual services on SD-WAN as a platform. That's the really exciting part. Security is the first of these services that runs on SD-WAN as a platform. Because of the strength of VMware, having led the world in virtualization and now having a strong position in network virtualization through NSX, this is where the marriage, partnership, acquisition of VeloCloud by VMware makes sense.

Can you give me a couple of examples of customer success stories?

Yes. I'll tell you about three out of our 2,000 customers. First: A retailer with several thousand sites had a network where they wanted to get voice services out to each of their locations. Those services were expensive to get when running only on their private network and many times they couldn't drive the private network to the ends of each branch they had. Voice services are very important in the retail space because people still want to call up their store and ask if the product is in. They want to have some human interaction. We went in and there was some skepticism initially saying we've had this network for 15 years. How does SD-WAN work, and who is VeloCloud? We had a partner in there. We went in, and within 18 months we had almost 1,500 sites deployed. That was early on. Now within 18 months, we can get 5,000 sites deployed. It's going at a much faster pace.

Examples (continued)

A company that has almost 3,000 sites in the U.S. and is expanding into Europe. Cost savings was not important. They were looking at how to automate and simplify each of these sites. The company does storage [like storage units]. They put in SD-WAN because storage locations are unattended. They can't afford to have IT at these locations. They wanted it automated so they could have locations up and running quickly. They looked at SD-WAN and it seemed like a no-brainer to them. We went in and within three months we had covered several hundred of their locations. We had a partner in there, a small partner, and this is a large enterprise. The partner worked with VeloCloud, and was providing logistics services, even doing some support. They're an integral part.

Examples (continued)

A service provider in Australia called McQuarrie Telecom did the largest SD-WAN deployment in Australia: 500 sites for a childcare provider. Families pay $100-$200 a day for kids to be in childcare, and the parents want to be able to look at and interact with their child during the day, and this is very expensive on a private network. We deployed SD-WAN to them and it's unbelievable how delighted those parents are. All 500 sites, and now company is expanding internationally. Parents can look at their kid during the day because we support media so well.