Charles Giancarlo hasn’t looked in the rear-view mirror since he left as Cisco’s No. 2 man in 2007. Back then Giancarlo was seen as heir apparent to Cisco CEO John Chambers, but Chambers didn’t officially retire until this summer.
These days, Giancarlo sits on numerous boards — including that of Arista Networks, which is in a legal dustup up with Cisco — and has founded several companies. He was early into the SDN game as an angel investor in Big Switch Networks. Most recently, SD-WAN startup Viptela announced that Giancarlo joined its board of advisors.
Giancarlo brings a long list of achievements and executive experience to Viptela’s board, including his work at Cisco as executive vice president, CTO, and chief development officer from 1993 through 2007. His team created the first IP telephone system and telepresence system and, in 1999, the first Wi-Fi product ever sold.
There seem to be a lot of companies in the SD-WAN space. Why did you choose to join Viptela’s advisory board?
One [reason] is that I really like the management and engineering team. That goes a long way with me. They’re extraordinarily talented, and I think they understand the business quite well. Two, their solution, I think, really displays a strong understanding of enterprise customers and what they’re looking for in a next-generation WAN access solution for their branch offices. Those include the need for security and privacy, the ability to use lower-cost Internet pathways to the current infrastructure, and the need to have multiple methodologies for backup in case primary lines fail, etc. I think they’ve done a really good job of understanding that and converting it into a solution.
I think it sorts itself out in the market. I think in the SD-WAN space, the technology still has a lot of unknowns, and it has a lot of opportunity for differentiation and change in the market. Every vendor tries to feel their way in there and tries to influence their customers’ choices toward their solution. Cisco is no different. They’re going to try to influence customers in their direction, including trying to set up standards that are preferential to themselves, or make modifications to their product that accomplish some of what the customers are interested in.
It’s a competitive market out there, but I do think that there also opportunities for companies that start with a fresh slate and a fresh way of looking at things. Viptela doesn’t come with a lot of the 20 to 25 years of historical legacy that Cisco has to deal with. For customers that are looking for a new environment based on lower cost and higher security, I think it’s a fantastic solution. The rapid adoption of their technology, by large enterprises, and also by Tier-1 carriers like Verizon and Singtel, highlights that.
Do you think there are too many SD-WAN vendors now?
Too many for whom? [Laughs.] Again, early in a market — and I’ve been fortunate enough to see a lot of new markets come along — there’s always more competition in the buzzwords than in the actual products, so I think you’re seeing the same thing here. SD-WAN, SDN, and virtualization, these are words that convey a lot of hope. As soon as these words become popular, every vendor says they have something there. Over time, the companies that actually have products differentiate themselves from those that are just using the words to apply to things they already have. So, look, I think there’s a lot of competition for SD-WAN right now, but I think there’s less than a handful of companies that actually have something in that space and Viptela is one of the leaders.
Now that SDN has been around for a while, where do you think it’s headed?
I would say there are three places where SDN is most being felt, albeit still relatively low-volume. There are two places in the data center. One is in what I think was the original meaning of the phrase, which was a virtualization layer, if you will, for the networking space. Oddly enough, I think that’s an area that has been less developed than some of the other areas, but you have Arista there, you have Cisco there with its Application-Centric Infrastructure, and you have Big Switch. You really don’t have very many others playing in that space.
Now, you’ll say: What about VMware? VMware is playing in a second space in the data center. So the second meaning of SDN in the data center is really about the orchestration of hosts with storage. The network plays a secondary role; the network is connectivity. But frankly, they [data center management] don’t have to change configuration of the network in order to change the configuration of hosts and storage. It turns out that for most people, the thing they deal with the most often is upgrading and moving hosts and needing to change their storage targets. What you’re seeing with the VMware purchase of Nicira [which created the NSX product] is a much greater focus around data center orchestration, around host VMs and storage, than you’ve seen around modifying the way the network switches behave. VMware’s NSX also includes, by the way, the orchestration of network services such as load balancing and security. That’s the second meaning of SDN; managing hosts and targets.
The third meaning gets us back to Viptela, and it’s a more focused area within SDN, as it’s now referring to SD-WAN. Really, what that’s about is creating a greater centralization of both configuration and management and the ease, if you will, of managing very large, extended wide-area networks, which traditionally is a complex, disaggregated portion of the infrastructure.
You mentioned Arista. From your perspective on Arista’s board, what is your take on the overall infrastructure market? What do you think will happen over the next few years, and what does it mean for the incumbents?
It’s a great market. Growing very well — lots of actively and lots of focus by the customers. Good focus by the customers means they really consider what vendor they chose from, so it’s a good, competitive market and one that new vendors can penetrate, which is what Arista is doing. I think the reason why Arista is doing as well as it is, is because the two prior dominant vendors in that space, Cisco and Juniper, really didn’t deliver the kinds of products that the customers were looking for in the timeframe that the customers were looking for. So it really gave a new vendor like Arista an opportunity to penetrate.
The same sort of thing can be said about the SD-WAN space. I think because WAN access routers have been a declining market over the last several years, both of those vendors [Cisco and Juniper] kind of missed the opportunity to bring out the new set of products that Viptela and others are going after. It gives new vendors the opportunity to penetrate the space.
I can’t really speak for the company. Yes, I’m on the board, but typically, of course, the CEO speaks for the company.
What are your thoughts on NFV?
NFV is really about several things. One of them is the carriers trying to create a more competitive environment for the products that they spend the most money on. Second, it’s about the carriers attempting to be able to interact with the products they buy from the vendors in such a way that they can add more value. In other words, a more programmable environment for the software they operate inside of the networks. Three, it’s about reducing the number of core routers in the network and relying more on switching. NFV is really about those three things. I think it’s going to take a long time for that vision to be fully fleshed out for lots of reasons, including that it is not necessarily in the vendors’ interests.
There’s not a broad landscape of talent in the world for producing carrier-quality software in this area. And frankly, hardware is required. It’s not so easy to replace some of this stuff with software. If you put it on processors, it’s not necessarily cheaper. I think this is going to take a long time to realize the vision there, but the work will continue.