Cisco Charts New Course

Q&A: Cisco CEO John Chambers describes how next-gen data center will affect business.

As Cisco Systems continues to push further into the data center, CEO John Chambers is looking to make Cisco a strategic partner to large IT customers there. At Cisco's annual C-Scape industry analyst conference in December, Chambers went so far as to declare Cisco's intent to be a leader not just in the networking and communications industry but also in the computer industry itself. Senior Editor Paula Musich spoke with Chambers about Cisco's next-generation data center initiative and how Cisco intends to achieve it.

Cisco's role in the data center today is as a supplier of SAN [storage area network] switches, network switches and some security hardware. How do you see that role evolving over time with your Data Center 2.0 vision?

What we see is that the data center vision is changing rapidly: It's changing around a concept of any device to any content-processors, storage, applications, with proper authorization anytime anywhere. And it's moving from [the idea of] a person to a machine, to many to many. It's what we call the second phase of the Internet.

How would you characterize the first phase of the Internet, compared with the second phase?

In the first phase, you had a person to a content-entering orders online, employee self-service, customer self-service and the business model change [around] core versus context, where you [have a] contract manufacturer do the manufacturing heavy work for you. Our expertise was not how to manufacture, but how to get the quality. That set the direction for the industry for a decade, and Cisco led in the implementation of that by listening to our customers but also catching the market transition. It wasn't about huge, complex data architectures. It was about really capturing the power of the Internet to allow people to do their jobs one to one or one to machine.

This next wave is about many to many, and it's around collaboration. It will result in a replay of the productivity [gains] we saw and predicted in the mid-'90s.

So the vision of the data center is that we want to be able from any device to get access to any combination of servers, storage and apps and make it completely transparent and secure to the user. For that to work-especially if you think about it as many to many and not one to machine-it is all about networking.

The network will become the platform not just within the data center but all the way to the home, to where you will not care if the Duke basketball game or my wife's "Desperate Housewives" is stored on the set-top box, on the PC, the TV behind that [or one of] multiple DVRs in the home. I don't care whether it's in the point of presence with the service provider, whether it's in the content generator down in Hollywood or somewhere out in the World Wide Web-I just want it to work. And I don't want to have 15 copies just to find that I made a mistake and overrode my wife's favorite program or vice versa.

And it has to start in the data center where, at least in theory, bandwidth is free and, at least in theory, you can touch any of the servers or storage devices or any of the apps. If you can't do that in the data center, how do you take it all the way out into the environment where, basically, it will be driven not by personalization but by multiple people working toward a common goal?

So the complexity goes up. It was a lot easier when it was a person to a device or one person to one other person sharing multiple pieces of information.

Video will be the killer app in the Internet. And it obviously will be in different forms of unified communications-Web 2.0, video blogging. I do no text blogging because I'm slow and I can talk 200 words a minute and really get my message across. So my blogs are video, and [Cisco's] TelePresence [virtual meeting system] in and of itself puts loads on our networks-probably incremental 200 percent growth per year.

So you're about to see what we believe will be the next wave of productivity [growth]. There's something we would like to call the "Cisco TelePresence network effect" or "Cisco's TelePresence law" that's the classic: The power of the network is the number of nodes squared, [where] 10 units is the power of 100, and 100 units is the power of 10,000. You'll see us lead, but you'll probably see some of our customers like a Procter & Gamble [scale] the number of units they'll use, and they'll change how they interface not just among themselves on this collaboration-where access to anything in the data center can be shared with any user-but they'll take it all the way to their suppliers.

When I think of Tele??íPresence, I think of it as an expensive product that will be sold as onesie-twosies in the executive suite of Global 2000 companies.

It's the reverse. I would never have developed it if it were primarily for CEO-to-CEO communications. The CEOs grasp the change, and they grasp the ease of use. It's as simple as a phone-actually, simpler. ... Now think of how the data center is core to delivering that-especially as video becomes more and more a key part of storage and apps. Think of the security implications. It's an architecture, not a product. People who approach it on a product basis [and who] are trying to integrate products that were not designed to work together are going to have trouble-especially if you're trying to inspect every packet on the network over any combination of networks.

It's a combination of data, voice and video together off multiple applications. It isn't anymore about servers, network virtualization, application virtualization and what do you do about interfacing this between data centers; it's about how you approach this total market. So the market transition to us is that the network will become the platform for enabling all forms of communications in IT. And any person can put on their PC and drive it from there with any database architecture they have.

So you start in the data center because that's where the money is available? And as the economies of scale drive down the cost of technology, then it moves out to a broader audience?

The reason you start in the data center is because broadband is free there. So your bandwidth is relatively free, and all your servers are available without delay and all the storage is there. So if you can't do it in the data center, where you have IT pros pulling it all together, you sure as heck can't do it combining points of presence all the way to the home or to my various business unit offices.

Do you see any potential roadblocks in the communications landscape that could derail that vision of going all the way out to the home with TelePresence and this kind of collaboration? I'm thinking specifically about net neutrality.

For me, the question with net neutrality is: What are we trying to accomplish? We're trying to get broadband to every American at an affordable cost with at least two to three technology options that they can choose from. So the primary goal is how do we get it built, and then later you can ask, well, how do we want to splice it in a way that allows each of us to gain access to a certain amount of broadband as part of a public utility?

But as a [service provider], if you have to deliver that free of charge to your competitors, or at a charge below or at cost, and you can't guarantee your response to me as an end user, then all of a sudden you have a problem. I [as a business user] will pay a premium to make sure I am able to get my applications. If I have to compete with a kid down the block using a computer game with 1,000 other people and I can't use TelePresence for an $8 million sale, all of a sudden I have a problem. And I'm putting TelePresence into my home this month. So you begin to see these are not separate. They have to combine together.

We tend not to say, "Here's where the industry is going, and jump." We tend to listen to our customers. I wish I could tell you I bought Scientific Atlanta because it was really smart. It was Time Warner and some of the other big service providers saying, "Go buy it." We bought WebEx because of a phone call from my top business person saying, "John, they're about to be bought by one of your competitors, but this is a company that's perfect for you." We thought [the competitor] was Microsoft, but it turned out not to be Microsoft. We got that phone call on a Thursday night, and we bought [WebEx] one week later. Everything was done.

We used a lot of these principles of unified communications and collaboration and Web 2.0 to pull that off.

Market transitions often are challenging to the incumbent, and our approach in the data center is to say, "This is inevitable; we're gonna lead."

And so you'll see a series of announcements over the next 12 months, one after the other, on data center switching. But it's really about data center virtualization. You saw us [take] a minority interest in VMware. That wasn't a subtle message to the market. That says we're going to play there. And we'll partner tighter than anyone else with the leaders in this.

Many of the CIOs will tell you there's been a lack of innovation in high tech for quite a few years. If you think about the innovation that occurred in the '90s, that was about transforming business models, supply chain, transforming the way you interface to your customers, the way you order and deliver online, your ability to communicate with others through e-mail first, then text and now video.

Now you're starting to see the next wave of innovation. Most CIOs are pretty hungry for this. Because, guess what? Their budgets haven't grown very much. And part of the reason they haven't grown is because productivity in the U.S. has been relatively stagnant for three years-in the 1 to 2 percent range, which is where [former Fed Chairman] Alan Greenspan believes it will stay.

But in the early '90s, the foundation was laid for the 5 percent productivity growth that occurred in '96 and '97. We called this one cold. In 2002 and 2004, we got years where productivity stayed at 4 percent. Now it's been back down again, and you're a CEO trying to make decisions and you're saying, "I can acquire a company, I can buy back my stock, I can move into adjacent markets or I can spend money on IT. If I can only get productivity of 1 to 2 percent a year out, I'm not going to spend a lot of money on IT." So what's missing is innovation. And what this data center virtualization enables at first is cost efficiencies, consolidations. But over time, it is innovation for applications, which I think will be built largely around collaboration or Web 2.0 concepts and directions. So that's our vision.

If you think about it, any device to any content-especially to any applications to any processors-that's what Cisco does. And it's about intelligence in the network, not just dumb pipes.

Can we expect to see partnerships between Cisco and other server virtualization providers such as Citrix Systems and Microsoft? If so, how soon?

We don't share our moves prior to making them public. I think you will continue to see us active both from an internal development, a partnering and acquisition role across the whole data center. We have an expanding relationship with Microsoft, and we have similar views of how the data center might evolve. And so I'd be surprised and disappointed if you didn't see us move with Microsoft on some opportunities, even though we might compete in some areas, like unified communications. That's really the future.

The days of complete friend or foe are long gone. Within that, EMC is the one that is aligned closest with Cisco with very little overlap. But we clearly want to partner with all the players-the IBMs, the HPs, the EMCs, the Microsofts. We are one of the few companies that have traditionally done that, and I don't underestimate that it's hard.

Check out's Infrastructure Center for the latest news, views and analysis on servers, switches and networking protocols for the enterprise and small businesses.