Judy Estrin has been an innovator and catalyst behind industry changes for many years. From 3Com's acquisition of Bridge Communications 11 years ago to Cisco's recent purchase of Precept Software, Inc., Estrin's companies have always had something the Big Boys want. Perhaps it was Estrin herself? Network World Senior Editor Jim Duffy recently spoke with Cisco's visionary new chief technology officer about where she sees Cisco going.
Q. How does Cisco make a decision whether to partner with a company, acquire that company or develop internally?
A. I would say sometimes it's a very carefully thought out plan and sometimes it's opportunistic. It's a combination of ... there are certain areas where the company looks at (an acquisition and asks) do we have that core technology, should we go look at partnering or look at acquiring? There are other places where we may come across a really great company and even though we have some of the expertise we may choose to acquire. There are places where for time-to-market reasons we may acquire. But then there's other areas where we may look and say no, this is really a core competency that we have and let's go develop it. It's all about understanding what our strengths and weaknesses are and where we want to own a technology vs. partnering. Pretty much the same way that most companies make that decision.
Integrated into the culture here is we're moving too fast to do everything ourselves. Looking to somebody else to help us via a partnership or an acquisition is not a sign of weakness, it's a sign of strength. It means that we're looking forward. I think that's one of the fundamental differences that I've seen at Cisco from some other companies.
Q. If we could break Cisco down by lines of business - small/medium, enterprise and service provider - how do you see each evolving? What technologies can't Cisco currently provide for those segments?
A. Most technologies have applicability to almost every segment. At a technology level, that's a lot of cross technologies. The differences in those markets tend to be how those technologies are applied and how they're implemented.
In terms of what technologies can't we provide today, that's a hard question. It depends on where does the network stop, what is the customer looking at. If you're looking at an end-to-end enterprise network, Cisco does a pretty good job of providing those pieces and then works with partners on the systems integration side, on the desktop or workstation side. In the service provider, which is a newer market, I think we are moving in that direction. Historically, one of the big areas has been the network management - the billing, accounting, and I think you will see more coming from us in that area. I think we either are providing or have plans to provide or have partners to provide most of the services that those people are looking for.
Q. How deeply will Cisco get into voice? Are we going to see Cisco branded central office switches or PBXs?
A. Cisco is very serious about this area. As you see more and more voice capability migrating to packetized voice as opposed to circuit switched voice, Cisco has a clear advantage moving into that market. The first phase of voice/data integration is really integrating it in the infrastructure with PBXs and data.
The second phase of that is truly integrated applications where you have PBX technology built on packet-based nets. I think you will see Cisco be a big player in both of those areas.
Q. Do you see any particular technology winning out in the Internet core, either ATM or packet-over-Wave Division Multiplexing?
A. I think what you will see is IP everywhere at the desktop, no question about that. I think in the core you're going to see combinations of ATM and IP. I think there initially will be IP over ATM. Then you're going to see IP directly on SONET, and then eventually IP directly on WDM. But I think for a very long time you're going to also see some set of applications that will continue to run over ATM over SONET and ATM over WDM. I don't believe that ATM will go away in the foreseeable future. So you'll really see ATM and IP needing to sit side-by-side. I think the one thing that is absolutely clear is ATM is not a desktop architecture, that ATM will be part of the core.
Q. When I hear or think of policy-based networking initiatives, like CiscoAssure, I'm reminded of VLANs, which didn't really take off. Will policy-based networking become the next VLANs? Will users find that the benefits of policies aren't worth the time or cost of implementation?
I don't think (they're the next VLANs) because we need some of these internetwork services like QoS, like security to do what we need to do. So therefore if you have these capabilities, the policy servers really make it possible to use those. VLANs were more a case of how you organize your network, and so there were ways around it. And the fact of the matter is, if you have an application that needs QoS, you need a way to set the policy on the QoS.
I think VLANs were driven by a, 'How do I find a proprietary lock-in to keep my customers coming to me?' (philosophy). That's why they failed. I think things like CiscoAssure is driven by a need the customer has that we need to find the right way to fulfill.
Q. So CiscoAssure won't be a proprietary offering?
A. I'm not saying that there aren't proprietary pieces to it. Sometimes companies have proprietary pieces because standards don't exist. In the early stages of the market, that's the way you do it. And then you interface to other people with APIs. But you don't want to come up with a feature just to have something proprietary. If you look at Cisco's early technology in routing protocols, they have some proprietary protocols because the standards didn't exist at the time. The customer needed that functionality. So you look at a customer need and you say, 'Okay, are there standards that exist to solve this need?' If there are, you use the standard; if there are no standards then you need to come up with some proprietary technology to use for that. But as long as (a technology) is driven by customer requirement, it has a much higher chance to succeed than if its driven by something the vendors sit back and say, 'Ohhhh, how can I come up with something new to lock the customer in?'
Q. And VLANs were more vendor-driven than customer need-driven?
A. I think so. I was not directly in the networking industry when VLANs were around. But I remember reading all of the stories about them and kept saying, 'Now why is the customer going to do this?' VLANs never made all that much sense to me.
Q. So users are driving policy-based networks?
A. Yes. Users are driving applications that require policy-based networks.
Q. So you think moving to a policy-based networking program like CiscoAssure would be in the customer's best interest?
A. No question about it.
Q. Now if I could get your thoughts on some recent industry events like Tellabs' acquisition of Ciena... What's Cisco going to do for WDM now?
A. Work with Ciena. We have a partnership with Ciena and I don't see this changing it at all. I don't see Tellabs as a direct competitor of Cisco. I think we'll continue to have a strong partnership there.
Q. Lucent's new PacketStar routers: threat or menace?
A. Clearly, Lucent moving into the data space (will make Lucent) our key competitor. In some ways, John Chambers likes to say we've left our traditional competitors behind, although you never ignore (them), whether looking over your back or looking forward. But if you think of Cisco, looking over our back now is at 3Com and Bay; looking forward is at Lucent for two reasons: one is we're moving into the telephony space, taking advantage of the move of telephony over packets, which is something Cisco can really lead in. That pits us against Lucent but also Lucent has very clearly stated that they're interested in the data market . Lucent is clearly targeted at it. So, no question that anything Lucent does is something we take seriously and we watch. On the other hand, what they don't have is the IP expertise that Cisco company has.
Q. Sprint's ION network: What is Cisco's role in that?
A. Cisco is one of the major equipment providers. We view it as, again, another example of just the industry moving forward. Our partnership with them is a tribute to our thrust in the service provider space and how we're moving to partner with service providers. It's a good thing for Cisco and a strategic thing for Cisco.
Cisco to buy IP multimedia vendor Precept
Network World Fusion, 3/12/98.
Geek of the Week: Judy Estrin
Interview with GeekWeek, 6/1/97.
Judy Estrin Spots Trends Before They Emerge
Entrepreneurial Edge magazine, 1/98.
Take it from the top: An interview with John Chambers
Network World, 10/6/97.
Kozel: Single network coming
Cisco exec discusses the future. Network World Fusion, 10/8/97.
Cisco stacks up with new products
Aimed at small, remote offices. Network World, 1/23/98.
Cisco: Impenetrable force?
Network giant has all the bases covered, or so it hopes. Network World, 4/20/98.
Cisco details optical networking strategy
Network World Fusion, 4/21/98.