Take it from the top: An interview with John Chambers

Cisco Systems, Inc. CEO John Chambers isn't in the limelight as often as Microsoft Corp.'s Bill Gates and Intel Corp.'s Andy Grove - the leaders of the other two companies in the so-called Wintelco triangle that rules the computing industry. But Chambers is no less outspoken and forceful than his cohorts, and he's not at all shy about proclaiming Cisco's dominance over the networking industry. In an interview with Network World Editor-in-Chief John Gallant, Chambers explains how Cisco plans to maintain its control and talks about the threat from a new-generation of routing startups, the evolving competitive landscape and Cisco's slow-start in Gigabit Ethernet.

Q. One of the things that is mentioned quite often these days is the so-called Wintelco triangle - that Microsoft, Intel and Cisco have significant control over the industry. Do you think that triangle really exists?

A. One of our top priorities this year is strategic alliances. And let me start with the candid discussion that almost all strategic alliances fail and most of them don't have any more value than just a marketing scenario. My definition of a strategic alliance is very simple. It makes a lot of sense to our customers, they benefit from it, they understand what you are trying to do. Secondly, it will result, by the year 2000, in $500 million to a billion dollars incremental revenue per year. Third, it's a competitive landscape change for both partners. As basic as that sounds, that's really how I would define the strategic alliance. We have, in my opinion, a strategic alliance with Microsoft and Intel, but also with Hewlett Packard, GTE, KPMG, and you'll see us fill in two or three more points.

Almost without exception, they were customer driven. Many of our customers have standardized on, or are going to standardize on, Microsoft. While you might argue about whether there is some concern with doing that, you can't deny where the majority of the enterprise environment is going to go.

As you begin to look at the advantages to the customer in terms of common security, common directories, addressing the cost of ownership issues - there are huge advantages for us to work with Microsoft. The same scenario applies to Intel. Same scenario applies to Hewlett Packard, just with different component parts. In terms of our top seven initiatives for this year, strategic alliances is number four. There are major advantages to an Intel working with Cisco and Cisco working with Microsoft.

If you go back to the strategic alliances that worked over the years, they've been very rare. Intel and Microsoft grew up together. If they had not grown up together, that alliance might not have occurred. Perhaps HP and Canon. But then after that, I'm really hard-pressed to say who has this true billion dollar a year type of [partnership].

As you go into it, you've got to say what you want to do differently to make these work in ways that others haven't. That includes everything from changes in our organization, to identifying what's really imortant and getting both groups to focus on what really allows for the competitive paradigm change.

The problem many alliances have is they focus on a transaction or two. I've never had a single transaction that I'm aware of that is 50/50. So if you do a lot of transactions together - where perhaps my partner will get 70% of benefit one time and the next time I might get 80% of benefit and it all adds up to 50/50 - that's the way it will work. Secondly, you have to get your whole company to turn on it, and drive it down through the whole company. If only certain segments of the company move with you, you will never fulfill the opportunity.

A third key is we restructured our company. I have Charlie Giancarlo, who did our mergers and acquisitions and used to report to Ed Kozel, reporting directly to me. We gave him strategic alliances, reported the partner engineering group to him. We gave them business development resources and then made [partnerships] one of the top four priorities in the company. We did it in classical Cisco style, much like we did with acquisitions.

People forget the first [acquisition] we did, my stock went down the day we announced it. They thought we'd lost it in terms of the frugality and not understanding that acquisitions just don't work. Well, the reason it did work for us is that we looked at the common vision, short-term win, long-term win and chemistry, as we discussed before. We rewrote the textbooks on how to do that and it worked beyond our wildest imagination in terms of the benefits that it's given to our customers and to Cisco. We think strategic alliances have the same kind of leverage and the same type of potential, but they're even more tricky than acquisitions.

Q. I'm asking this in a different sense though, beyond specific strategic alliances. There's a sense that where Microsoft has a certain level of dominance within the software industry and Intel controls the semiconductor industry, Cisco is in the same position within the network industry. Do you agree?

A. If you look at our market capitalization, which is the financial community's expectation of the future, we are the size of the next 10 competitors combined. The financial market has already voted that they view us that way, if we execute properly - and time will tell whether we do or do not. We have a potential revenue and profitability impact over the next five or 10 years that all of our competitors combined do. I think we have the chance to shape the future of the industry. You will see us do that in a kinder, gentler way than some of the leaders have in the past, such as IBM. That's where both our partnering philosophy and open standards play a much different role than some of the people whom we've met in prior generations of technology evolution.

Q. What about in the sense of technology evolution? Do you feel as though the technology stance that you have today is as protected as, say, Microsoft's? Do you really have a dominance over the industry that would be difficult for someone to dislodge?

A. To put it in proper perspective, our goal is to make [Intel CEO] Andy Grove look relaxed in terms of his own paranoia. We worry about everything, I'm not secure at all. I can tell you about market segments that have started up that we're is worried about, I can tell you which are the good start-ups and who might be new competitors coming into the market from other areas. Maybe 'I'm not secure' would be the wrong choice of words, but I'm extremely paranoid and I think that's very healthy for a company.

Having said that, do I believe that we are the only company in the communications industry that controls this future? Yes, I do. So it's a matter of how well we execute, how we stay close to our customers and avoid the prior mistakes of companies who got into this position, or close to it, and then misstepped.

Q. Most people tend to look at that kind of control as a negative. What's the benefit of having that kind of control for your customers?

A. Well, let's talk about a huge trend that is occurring in the industry at a pace faster than anybody dreamed - one that I did not even think was occurring until months ago. Customers are going to fewer vendors in the network. Consider a PC that goes on one of your reader's desks. The cost of that PC is only about 25% hardware and software. The real costs are in systems administration, the help desk, how do you do software releases, how do you do troubleshooting, etc.

The same is true in the network. The majority of costs are in the support structure in terms of systems integration, systems administration, the communications lines, the switch lines. Our networking component part of it is only about 25-25% of the total cost.

If you have not only fewer vendors in the network but perhaps a strategic vendor in the network, and if you could bring in other partners that take it beyond the network such as Intel or Microsoft or Hewlett Packard - then the payback to the customer is huge. So what you are seeing in the enterprise accounts, in the service provider accounts and now the systems integrators, is not only fewer vendors in the network but very serious consideration given to a primary vendor of the network, period.

And our win rate here is extremely high. It started in New York with two financial customers saying this is the way that they wanted to go back in January 1996. I knew both of them well and I actually argued with them in a fun way, saying that I did not think they would want it to be like IBM of 25 years ago. And they commented back to me that that's exactly what we want, we want somebody in this new technology that not only has tremendous productivity for us in the future, but may even determine our growth or possibly even our survival. We want somebody who can be a true partner, but not make the mistakes of IBM over the last 10 years or perhaps some of the other industry leaders of today.

They said Cisco already has the number one or number two products in the majority of the areas they are looking at. With [Cisco's] combination of do it yourself, partnering and acquiring, we think you are going to stay in that role and move into areas like token ring or network management where you haven't already gotten that position. So they were way ahead of themselves in terms of perception of the market. By the end of that calendar year it was almost 50% of the Fortune 1000 considering it.

And I can honestly say today, I haven't had one key executive say they would not actively consider that. Ninety-eight percent of the customers say, hey, I'll actively consider that, and then you have telcos, who traditionally differentiate themselves even by the type of vendors they had in the network, and systems integraters, who differentiate themselves by making the various component parts work together, saying this really makes sense for me as well. It means that you are on the front end of a wave that has tremendous advantages if it's executed properly.

Q. And what is the risk of this?

A. The risk is twofold. First is that we misuse the relationship, and second is that we don't keep up the pace in terms of our product leadership.

Q. How will you know whether either of those risks are materializing?

A. The issue is very simple: if you are not gaining market share or sustaining a substantial part of the market share, that's a very direct message that you are not maintaining the advantage. Second, you want to constantly compare yourself to both the individual point products as well as the total solution. And in that area, the customer will tell you, if you listen to them.

One of the reasons we focus on the customer so hard is that I tie the compensation of every manager to customer satisfaction - really listening to customers, walking our talk. We are the only company of anywhere near this size that does it. I review every critical comment in the world every night myself. Not that I fix it, but if I pay attention to it, everybody else does.

So I guess the answer is that we'll know when the customers tell us. The market will tell us in terms of market share and growth. But as long as you do it with open standards, which we are, and as long as you are not dealing with an operating system for which people have to write applications - which is your real cost of switching over - it's no more difficult to change any element in a Cisco product line then it would be for another vendor.

So the risk to the customers - in comparison to an operating system - is much lower and the potential benefits are huge.

Q. Do you see Microsoft as an ally, or a threat? I know you have a project that you are working with them for directory services, and yet, they've been fairly vocal that this Steelhead routing/remote access product, is a competitor to some of your products.

A. If you ask Bill Gates or [Executive VP] Steve Ballmer both of them would be very articulate in saying Cisco is a partner, period. We are not going to compete. We went in [to our partnership] knowing very well the decision they made two-and-a-half years ago to develop [Steelhead]. They offered to work with us for a fixed price and when we turned it town, they went with Bay. So we knew that product was under development, it would be an issue there.

The issue is: is it part of a standalone capability, or is it really beginning to get network functionality? You have to have in the Internet appliance, be that a PC or whatever, some basic functionality that allows you to interface with the network and that's going to occur whether they did it with us, on their own, or with somebody else. So that does not surprise us nor in a big way concern us.

There was some initial positioning by some of the people in Microsoft and, candidly, by some outside people who want to stir the pot about how this would be directly in Cisco's space. But if you look at their support structure, their distribution structure, it's not in their best interest to do that and I think they clearly understand it.

Second, if you really look at the benefits of us working together, they far outweigh that one area of problems. I think we have a pretty firm agreement they won't move past that area.

Time will tell, but I believe that both Bill and Steve are very committed to partnership and they understand the value of partnerships in a way that perhaps they may not have or would not have been willing to a couple of years ago. No one, regardless of size, can go alone anymore and the companies who do, regardless of size, will be left behind.

It took one meeting with Gates and one meeting with Grove for them to understand the concept [of partnering], the second meeting to say 'this really makes sense'. I talked to IBM for six years about how you do this in the market and the benefit to their customers, and it just speaks to the leaders in the industry today versus perhaps some of the leaders in the past and how different the awareness is. That's not to say [IBM CEO Lou] Gerstner is not trying to change IBM and making good progress, because he is. But changing that culture is like turning not an aircraft carrier, it's like turning around Manhattan.

Q. Let's talk about the competitive landscape. Your competitors to date have largely been the other three of the so-called Big Four traditional data networking players. With 3Com's acquisitions, there's really a Big Two and then there's Cabletron and Bay. Are you more concerned about those companies still, or worried more about the Lucents, the Nortels, the other big phone switch companies that are getting into the data market? Who are the competitors going forward?

A. We've got our whole company focused on competitors by market segment and we learned from Hewlett Packard that that was the right way to do it. Once we began to develop a market-segment competitor focus, we started asking what competitors are doing in that area that customers like. And like Hewlett Packard taught us, as we segmented our market, they were different in every segment.

In the traditional market it was 3Com or Cabletron or Bay. Probably Cabletron first, Bay second, 3Com Third. But in ATM it was Fore [Systems]. In terms of low-end access, it was Ascend. In terms of wide-area switching, it was a Cascade or Nortel. More recently, in this area called interfacing the mainframe, our competitor is IBM.

By the way, as I travel, I ask how many people's mainframe transactions are increasing. Almost all of them are, usually 20% to 50% per year on the average.

So our competitors are different in each market segment. But as the customers begin to move to fewer vendors in the network and perhaps one single partner, then you've got to look not only at the market by product type, you've got to look at it by customer focus. So we now look at service providers, enterprise accounts, small-to-medium business and, beginning sometime in the next calendar year, the consumer. Our competitors are different in each of those areas.

In the service provider marketplace, it is some of the large players that you've said and it remains to be seen if Lucent and Nortel would be long term competitors or long-term partners. But it will be one of the two. It is new players like the Ascend/Cascade combination, if they make that combination work. Time will tell, because doing mergers across geographies is extremely high risk. But I think Mory [Ascend CEO Mory Ejabat] is a good leader, he could pull it off. But there are very high risks if they don't do it. They could be a key competitor there.

In the enterprise accounts, surprisingly enough, it's still Cabletron. They have the most direct sales force, and when you ask my field reps who they see, frequently it's still Cabletron in the enterprise account. Small to medium business, 3Com and Bay. To the consumer, it's pretty wide open to a whole bunch of different players and we'll go after that through partnering.

So the long way to answer your question is that our competitors are different in each market segment and they are different as we begin to go by line-of-business focus. Our philosophy has always been, in the customer's best interest, to partner and to have a customer benefit by partnering rather than competing, so we will partner wherever possible.

Q. 3Com made no bones during the acquisition of US Robotics about the need to stay with Cisco in terms of size. What do you think of 3Com today and the merger of 3Com and USR?

A. [3Com CEO] Eric Benhamou is an excellent leader. He's one of the very best in the industry and I like him and I trust him. However, I love beating him whenever possible. But I think [the merger] speaks to how rapidly we are pulling away from the rest of the industry. The fact that Eric would do a combination with a company of equal size, halfway across the country, speaks to how much risk he was willing to take on to attempt to stay close to Cisco.

I would argue that it was also as much a focus on Intel as it was Cisco in terms of modems, NICs [network interface cards] - the majority of 3Com's revenue, actually. Intel is probably a much more serious competitor to their future revenue strengths than Cisco is.

I think the financial market articulates it well. Even with [3Com/USR's] combined size, the market's expectation of their future is about a fourth of the business of Cisco - even though the combined revenues are fairly close. I would say it positions them perhaps better versus Intel or Cisco if they make it work.

Putting together two companies across geography is tough in terms of getting the efficiencies that you want, keeping out the politics, getting the future growth. Our fifth rule of thumb was geographic proximity when doing large acquisitions. We combined StrataCom and Cisco in 90 days. Now, we could have done some things better, but our manufacturing head count last year increased 7% when our revenues went up 57%.

That's because you combine the efficiencies of people who had common career paths. What people miss is that, let's say Adam [Adam Stein, a PR manager] is in PR in Chicago and I say 'Adam, don't worry you've got a future, you've got a job.' Is Adam very happy if my PR center is here [San Jose]? No. Nor do we get the efficiencies in marketing, manufacturing. So you have huge disadvantages by not acquiring the large acquisitions locally. All you are doing is acquiring engineering. The engineers run pretty well remotely. The closer the acquisitions, the easier they were to digest, and the more successful they were in a shorter time frame.

Q. Makes sense. Specifically, what are the weaknesses in the combined 3Com/ USR story?

A. Well, let me talk more about some of the challenges that they have. You've got your customers', your employees' and your shareholders' future to make it work. And mathematically, I've never seen a company that combines two equals end up in a number one or number two position in the computer industry. Think about that. Now, you mathematically combine them sizewise for a period of time and they were [number one or two]. Let's use Bay as an example, I remember when Wellfleet and Synoptic combined, they were bigger than Cisco. Everybody wrote about how they were going to be the one or two player and three years later, we are three times their size.

People tend to have a pretty short memory about the risk of mergers of equals and different cultures, and I don't think anybody would argue the cultures aren't different within these two companies.

The second challenge that they face is determining an innovative product strategy when their market is so spread. There is a lot of difference between products that are not commodities, but close to it, like modems and NICs, versus systems. They require different expertise and your ablilty to combine them in such a way that customers benefit is very, very tricky.

The third thing is that they are missing large product areas. Look at the major component parts where we are number one and number two. We are number one in high-end routers, they are not. We are number one in wide-area ATM switches, they are not one or two. We are number one in wide-area frame relay switches, they are not number one and number two. We are number two in security and firewalls - they are not in that category. And while they have reasonably good routers, they aren't number two in the router marketplace either. And in terms of interfacing to the mainframe, we have 73% market share. If you look at 13 to 16 areas of products, we are number one and number two in all but a couple of those, where they are missing more than half.

If you start talking about integrated solutions, a majority of their revenue is from NICs and modems and those aren't systems and they don't have the software that ties them all together - which is how you make changes quickly and how you get the products to play together in a way that allows the customer benefits to occur.

While 3Com is pretty good at acquisitions, I would argue we are better in terms of the business results we've gotten. Also, strategic partnering is mutual and strategic partnering can't be copied. In other words, there is only one Microsoft, there is only one Intel, there is only one Hewlett Packard, etc., and we've got them. If partnering has the benefit to the customer we believe it does, that is an exposure.

Q. Speaking of companies that merged across big geographic distances, what do you think of the changes over at Bay Networks under new CEO David House?

A. I think they've done a nice job in stabilizing the company. But I would draw a parallel to when they first combined. There was a lot of concern then. But about a year later, there was an upturn and all of the articles came out about how they turned the corner and things were going very, very well.

My comment to Andy Ludwick [former Bay CEO] at that time was that you will see waves and while the waves might be apart, you've only been through the first wave. Your real test will be the next 12 months because I think the hardest part of the acquisition is now to come. That is, can you generate the revenue, can you drive the business, or did you just stop the bleeding with the combination. Time will tell. The real answer is what Bay looks like two to three years out.

Q. Do you think a company that stumbles like that can come back in customer confidence?

A. I don't think a company that stumbles that badly can ever come back to the number one or number two player. We refer to it in dog years. One year on a calendar is seven years in a dog's life, so if you stumble over a time period of two years, which is really how long they stumbled, almost regardless of how well they execute going forward, they would have to grow double the rate of Cisco for - you can do the math as quickly as I can - I don't know, three or four years, even to get close revenuewise. We are growing dramatically faster than they are. They had a good last quarter, but we grew faster than they did.

It's a mathematical challenge, but also in terms of customer confidence. The customers understand that they've still got an East Coast company and a West Coast Company and our customers are picking strategic, integrated vendors and Bay, like 3Com, has loopholes in their product line.

How decisions are being made in customer accounts has changed dramatically in the last five years. I was with an account about three weeks ago, I won't mention the name, but they said 'I have every hub in the world'. They made decisions on their network and technology by geography. Since they were a decentralized account, they literally had every hub that had ever been made, to the best of my knowledge. I did not believe them and I said 'name them' and they named them by geography. I was shocked. But then shortly after that, people began to select a specific product worldwide because they realized the cost of ownership within a product set was major - how you get them to work, how you design your buildings. You could have huge efficiencies by selecting one hub vendor, a router vendor, a high-end router vendor, a wide area switch vendor, a token ring vendor, an ATM LAN switch vendor.

Following that, people began to realize that this was a strategic decision for them, because of the amount of money they were spending. But they also began to realize the productivity and the risk. That caused the industry to consolidate even more. Then you began to see people reduce the number of vendors in the network to just one or two or three. That's really driven by cost of ownership. I'll show you the charts, but you know the cost of ownership is your real issue here. And your time to market is a real issue.

Customers are beginning to say that they have to have Web-tone capability, just like dial-tone capability. That's the parallel that I'm drawing. It has to work all of the time, and they are realizing that with small companies, both to have a huge financial investment, but secondly getting a whole bunch of companies to play together to have the same Web tone is mathematically impossible. When there is a problem, the finger pointing occurs. But most important is the support of each of those various vendors, having to escalate through the problem determination, or having the expertise to be able to design it and make them work together.

So you are going to see a change there, where customers not only work with fewer vendors. A year ago, I would have been ecstatic if 10% of our customers would go with a preferred vendor or a single vendor network. It's over 50% today.

Companies like Bay have a real challenge there, because they do not have the product breadth to do that. They have good point products, and a large amount of their success over the last quarter or two has been because they've got a real good switch [Switch Node] which we clearly understand. They have a certain amount of demand there, so I'd be surprised if Bay does not have several good quarters in a row. But if you look out more than the 18 to 24 months range, that will determine whether they have a chance to go back to the top one or two.

Q. What did you think of the changes at the top of Cabletron with CEO Bob Levine stepping down?

A. I have a lot of respect for both [Chairman and COO] Craig Benson and Bob. I might have been different from others in the industry, but I always found him pretty easy to get along with, with the exception of one or two bumps and usually we hammered those out. Once a person no longer is having fun in their job, or they are not interested in the job as much, then it's time for a change. Sometimes that's also a scaling issue.

The other thing is that I think as the industry changes, bringing in people who really fit where the industry is going is a tremendous advantage. Using Cisco as an example, we are on our third generation of management, in fact fourth in many cases. The people who found a company, it's very rare that they make it through the centralized stage, then the decentralized stage, and very often people who are real centralized have a lot of trouble with acquisitions and partnerships.

Bringing in an outsider to be part of the senior management team, who has a different perspective and more experience with a larger operation made a lot of sense. I think the Ken Olsen's, the Ray Noorda's of the world are the real exceptions, and even they can only take the company to a certain plateau.

Q. Do you think this executive change will address this issue of personality or immaturity that Cabletron is wrestling with?

A. It will be a step along that line, yes. I don't think it will solve that in and of itself, but it was a logical move, Craig is a good business person. I have a lot of respect for him and I like Craig. I trust him, and that's really a lot to say about a competitor. I trust Eric [Benhamou] and I trust Craig. And Mory [Ejabat] and I are beginning to build that relationship at that level. He will, at times, literally bounce some ideas off of me in ways that require trust. I think our industry has moved out of the cowboy type image. You know, gunslingers, anybody who moves you shoot. That's terrible for our customers, the shareholders and markets because there gets to be a lot of confusion. You need to get the personalities out of it and focus more upon the underlying basics of business.

Q. There is a new set of competitors and I want to know whether you feel threatened by them. I'm referring to companies like Juniper that are promising new high speed routers for the Internet. Do you see them as a competitive threat?

A. Oh, the answer is absolutely yes. I view all start-ups with point products as competitive threats. Having said that, the market has changed. The customers are not making point product decisions in most situations. Many of them have already selected the vendor and to make that kind of change is painful to say the least. There has to be a huge competitive reason to change.

I would like to also remind people that that's the same question we got about Ipsilon 12 to 24 months ago. A great marketing campaign, a great promising campaign on where it's going to go, but I haven't been asked about Ipsilon in the last four months. There are many companies that create challenges for you at any given point in time and a lot is being written in the press about them.

But Cisco is almost too complacent, too nice, in terms of how we address some of our competitors' marketing claims. You'll find that we are not going to do that anymore. People make marketing claims on architecture and performance which we believe, at a minimum, should be challenged and at a maximum are extremely aggressive and questionable. You'll find us challenging them much quicker, not in a public way but with customers.

You are going to see us be more direct when people make marketing claims or functionality claims, put things in proper perspective in challenging those and giving the customer and the industry press the knowledge to challenge it. You'll see us challenge the start-ups more crisply, and get the data out to the market more crisply and take a strong position with customers on the performance.

Q. Why do you think Juniper has gotten this industry backing. (Juniper partners include 3Com, Nortel and Newbridge Networks, among others.) Are high-end customers looking for an alternative to Cisco?

A. Well, let's be realistic about who is backing it. They are either currently Cisco competitors or think they are going to be Cisco competitors. Let's not kid ourselves on that. It's quite a compliment to Cisco that these companies have said they can't compete against Cisco by themselves, to even have a chance, they've got to combine and provide input into anybody who might be able to challenge Cisco. You have people who are tremendous competitors, Lucent and Nortel, for example, funding this. These are companies who are very concerned about Cisco's competition versus them.

The next paradigm change for Cisco and the industry is the integration of data, voice and video. And we are moving very, very rapidly there. We showed an ability to move into the turf of players such as IBM and we've been very successful in that segment of the market. We've taken huge market share away from them, 73% market share.

For Lucent or Nortel not to realize that we could be either a very powerful partner, or a very powerful competitor, would be a naive position. We've showed an ability to move much faster on the product areas than they have, so I think it's a logical extension as the try to decide whether to partner or compete to make a move like this. It did not surprise me. I think it was a great marketing move by the company [Juniper] in question, but you have to ask how each of those partners really benefits. I think this is a vote of confidence for how successful Cisco has been. That is quite a nice compliment to us.

Q. What don't people understand about the story that companies like Juniper and others are pitching about the Internet? Will you have products that compete?

A. We already have a product today in the 12000 [Gigabit Switch Router] that we haven't even formally announced. It is already in the Internet in production at the OC12 level. It's already working and we've shown an architecture which can go up to 100 million packets per second if we execute it right. This is going to be a much tougher market to enter than people realize.

You have a majority of the service providers saying 'I need a strategic partner here'. Now, if we don't provide the functionality they need they'll have to choose another, look at other alternatives. But I'm cautiously optimistic about our ability to keep ahead of the curve - dramatically. Software is half the game here.

Q. Network managers hear a lot from analysts about Cisco being too router-centric. They say that you do this or that to protect your router base. Tell people what they should know about Cisco and routers for the future. What is the role of the router, and how will Cisco take customers to the next step?

A. Understand one thing about Cisco. We are technology agnostics. We really don't care. It's important to understand that. We do that for very selfish reasons. I've been with companies, IBM and Wang, that get stuck on a specific technology. If you're right, you are fine. If you are not, you've got a real problem. Bottom line, if customers want us to give them functionality, we are going to provide it. When we hear from enough customers, we'll provide it very quickly. We've shown an unbelievable ability to listen to the market. Our LAN switch sales surpassed our high-end router sales this last quarter. We are going to be the leader in each market segment, and we let the market determine who wins.

Now, let's take it a second step, and this is also an indirect answer to your question about Juniper and others. The days are over when you could sell just a router, or a wide-area ATM switch, or a frame switch, or a local LAN switch that does Fast Ethernet, or does Gigabit. You are going to see those products tie together logically first, and then physically over time.

Look at how our market has evolved in the LAN switching product area, the 5000 and beyond. We first did Fast Ethernet, and then we did ATM, we added routers, we added and designed a Gigabit product that fits straights into that. Our customers did not have to bet the future, their financial investment and perhaps the future growth of their company on which technology would be right for them. Anybody who says 'here is where the technology is going to be five years from now' without a doubt is overconfident, and probably wrong. None of us can project what is going to occur. Enterprise customers want a company that provides both an architecture and flexibility.

Switching versus routing has been an issue for five years, and as everybody found out, it wasn't one or the other, it was a combination. The lines between the definitions are going to dramatically disappear. So to answer your question in summary form, first is that it will not be one technology, it will be a combination. Secondly, we have no technology bias. We are going to be the leader in every major segment of technology. Number one or number two, or not competing. Partner with number one and number two. We've shown the ability to execute that again and again and again.

Third, we design products that not only logically work together, with a common software fabric, but physically over time will work tightly together. While it's always possible that it will be one technology that is the leader there probably will be a combination of them. So I think we are positioned very, very well. We've made a smoother transition in our product line than has ever been done in the industry and we've done it maintaining the leadership of the prior product line.

We've already made the transition, so I think it's a very good job of marketing by some of our competitors, and then perhaps not as effective a job as we need do in articulating where we are. But I'm hearing less and less of that. If you ask most knowledgeable Cisco customers, they understand that we are much, much more than a router company. In fact, our router revenues will be behind our switching areas in the future, just like our high-end routing is already behind our main switching.

Q. Let me ask you about Gigabit Ethernet. You bought into that early with Granite Systems but you are relatively late to market. Is there a problem with that technology?

A. Our view on the opportunity of Gigabit has not changed. We think with the price points, the functionality, it's going to be a major player in the market. Again, probably in combination, with Fast Ethernet . . . Gigabit feeding into ATM and routing. So it will be a combination of those. Secondly, when people say they've got a product, the question is, is it really a true Gigabit switch with all of the functionality of software?

We can not, nor will we, do the same thing a start-up does, which is basically bringing out a product and using Jeffrey Moore's analogy, Inside the Tornado, not worry about customer satisfaction at all when you first launch and let your customers help shake it out for you. You can't do that. Our customers put our products straight into the core of the network which runs their business, and so the expectations of quality are high.

We also made the decision not to design a standalone product. We designed it so that it was going to go first into the 5000 and then as a semi-standalone, probably in combination with other technologies such as Fast Ethernet and ATM. Unlike others, who might do it just for a stand-alone point product, we design it for working both logically and physically with where we have to go.

It isn't been any secret that [Granite Systems' founder] Andy Bechtolsheim's brilliance is in seeing where the market's going to go and getting there first. The density of our chip is the price performance advantage that we have. It's pushing the edge of technology and so when you boost the technology that much, your ability to bring it out as fast is a little bit slower. But if it's the price-point, and the design and the density we think it will be, we'll have a great advantage.

Q. Do you think customers care that you are six months later then competitors?

A. I think customers do care, but you have to understand one of the things we protect very carefully is our reputation. When we say we are going to do something, we deliver. And that's a lot different than almost anybody else in the industry. So we make a totally off the wall claim that we are going to become number one, in LAN switches when we don't even have a product under development, and six months later we are. Two years later, we have a 54% market share in the enterprise. We show an ability to really make things happen. We say we are going to move into Japan, and get the same market share in Japan that we have in the rest of the world. No American company has ever done that in hardware, and yet, three years later, we have 60% market share. I could use example after example after example of that.

So I think we have a tremendous amount of credibility and trust with customers that when we say we are going to do something, we do do it. That's not to say that we will always be the first or second, but we'll be in the first five, with products that work well and will be integrated - software, hardware integrated. It will be designed with stability and support in mind that other start-ups don't even worry about.

As long as we can outline to the customers where we are going and show them how we are going to get there, I think it's very acceptable. So if you watched how we did our 12000, we did not announce that a year ahead of time, in fact, we just announced it. But by the time we announced it we already had it in beta and in production sites.

Q. So there are no problems with Gigabit Ethernet? That technology is moving towards fruition?

A. Well, it's not to say that there are no problems. There obviously could be problems in front of us. But I'm comfortable with where we are architecturally and we are comfortable with our Gigabit strategy and how it ties into the total strategy. I would always like to have it earlier, but I think we'll be very competitive.

Q. About acquisitions. How many is too many? I mean, do 55 small acquisitions equal one overly large acquisition? At some point, you have so many different cultures, people coming from so many different backgrounds, who worked on so many different projects. Does it get too difficult to make the pot melt?

A. No. But remember, again, a large part of an acquisition's success is in your selection process. It's kind of like getting married. If you got married after one date, your hit rate would not be too good.

But if you knew what your criteria was and you spent a lot of time getting to know what you were after and spending time together, then your hit rate gets pretty high. We've shown an ability to do acquisitions with an effectiveness no one else in the industry has rivalled. And most acquisitions today are still failing.

If you want to know whether they've been successful or not, ask how many of their engineers are with the company two years later. Whether you pay $500,000 or $2 million per employee, you'll find that with most of the acquisitions in the industry, 30% to 50% of the people are gone two years later. Ours is single digit. So we've shown an ability to do it successfully.

We organized our company along business units and lines of business where you can do acquisitions for each business unit. You not only have the selection process down, but you do it in such a way that each business unit cannot do 10 or 20 but each business unit could do one or two. We have 15 business units so that gives you an idea of the skill we have. You'll see us acquire 10 to perhaps as many as 15 companies in a 12 month window.

That complements our $700 million R&D budget. It's actually $800 million dollars that we'll spend in the next 12 months on R&D. And we spent $6 billion on acquisitions in the last 18 months. That should give a customer pretty good feeling about our ability to keep up leadershipwise and that we are the partner they can look to for the long run.

Q. But wouldn't it also give a customer pause to consider whether you can continue to - and I don't know what the word would be but let's make one up - enculture everybody into the Cisco way and provide the same kind of service as before?

A. Part of it goes back to what your reward system is tied to. I think the best indication of the future is how companies have done in the past. I don't think anybody would argue that we've done this the very best in the industry in the past, and that's a good indication of what the future would be. We are very good on our selection process. The people want to work for the industry leader, they really do. That's why we keep them. It's an exciting place to work and you can make a difference here.

Q. Let me ask you quickly about the "CISCO-Powered Networks" marketing program. What is the goal with that?

A. Very simple. Customers are beginning to realize that dial-tone and Web-tone are the same in terms of expectation of product quality. And we are seeing more and more of our customers understand the advantages - in terms of reliability of the network, the cost of the ownership of the network, scalability of the network - based upon it being Cisco. And so they've asked us to help them articulate that in a way that really makes sense, that the end user customer can understand. That's the real benefit, both the message to the end user and what they can expect from it.

Q. So this is targeted to the carriers and ISPs?

A. This is carriers and ISPs, in terms of saying they are Cisco-Powered, but it's really targeted at the small and medium business or the enterprise business or the end-user consumer who says,'you know, I'd like to be a customer of that carrier or ISP'. Because if our customers decide to outsource a segment of the network, or outsource the whole thing, they are after somebody using a reliable system.

Q. Is it "Intel Iside" for the network industry?

A. It's a concept along that line, but much more complex. It isn't a product, a pin-point product issue, it's the whole network fabric of both hardware and software. But it was interesting, when we first came out with that, I did not get criticized by their carriers and service providers, I got a lot of questions on what we have to do to get to the same mark.

Q. Do you think the enterprise network buyer really looks at a branding like that? Do you think that kind of a campaign is a deciding factor in making a purchase?

A. The enterprise buyer does two things. First, if they are putting it in internally, they understand how important this strategic decision they are making is, as opposed to tactical. And second, if they are thinking about outsourcing part of it, they can outsource much easier than with somebody who already has Cisco if they've got Cisco. It gives them flexibility to choose among various players, or to later pull that back in if they want to for any reason, when it's a Cisco solution all the way through. A lot of the outsourcing that occurs, the enterprise customer will actually tell the outsource vendor, I want Cisco as my strategic partner.