• United States

Extreme changes

May 10, 20046 mins
Cellular NetworksWi-Fi

We talk to CEO Gordon Stitt about changes in the company, the competition and product direction.

Extreme Networks has changed a lot in a year. Since last spring, the vendor has launched a wireless product line, debuted its next-generation 10G bit/sec switch (the BlackDiamond 10K), and revamped its core switch software with a modular, Linux-based software operating system – ExtremeWare XOS. Extreme’s CEO Gordon Stitt talked with Network World Senior Editor Phil Hochmuth recently about the changes in the company, the competition and some product directions.

Extreme’s focus seems to be shifting from a core switch vendor to a more end-to-end approach. Why make this shift?

Different people have different perceptions of how much we’ve changed. I think it comes down to a change in the market. When we first started out, performance was the key thing. When you look back into the late 1990s, it was all about getting more bandwidth. But when things slowed down in 2001, 2002, there was tons of bandwidth installed. Now it comes down to a different set of issues. People are more concerned about security than they are about bandwidth. The whole talk about convergence is a big deal because it fundamentally can transform. We have taken a more systems approach. . . . This is very different from the speeds and feeds of yesteryear. Don’t get me wrong, performance still counts. You can’t go out there and say we run at 50 megabits and 100 megabits. You have to run at wire speed. But the first criteria isn’t performance, it’s ‘How do I solve this problem?’

But Cisco has a big head start being an end-to-end vendor. How do you catch up?

They are the end-to-end leader, and they do have a head start. But they do it in a very proprietary environment. It’s like the IBM AS/400; you used to buy your ERP applications and computers from the same company. You wouldn’t even consider doing that today.

Networking is just in an earlier stage, like that. Cisco is dominant in the end-to-end market. But 10 years from now, that will be an anachronism. You’ll choose best of breed just as you do today in computers and applications and communications devices.

Why was it important for Extreme to get into wireless?

I used to look at it and say, ‘Well if your network is already wired, why do [wireless]?’ The value in wireless is new kinds of devices being connected to the network. That makes it very different from wired Ethernet and creates different challenges. It’s not terribly hard to authenticate a wireless laptop today, but how do you authenticate a piece of medical equipment? Or authenticate a camera or a thermostat on the wall? If you look at it that way, I think wireless is very strategic to us. It’s going to happen in a big way. For me, wireless was a very strategic investment and critical to being able to continue to be the alternative to Cisco in the large enterprise.

If users will move to best-of-breed products, there are hundreds of wireless start-ups to choose from. Does that scare you?

There are hundreds of start-up companies in wireless. There are too many and there are too many by many times. But there are some interesting areas that people have addressed niches. I frankly think that a lot of people doing the wireless switches don’t really have any unique capabilities, and ultimately they will become subsumed by the incumbent vendors. We have a solution that may not do 100% of what every wireless switch vendor does, but it does 90% and it will do the other 10% in six months. We may be a little bit behind at any given time, but not for very long. Ultimately what’s key there is integrating with the wired network. If you look at voice over wireless over IP, you look at someone roaming and disconnecting his laptop. You want a single unified interface. People don’t want to have two consoles. If we’re going to go to an integrated voice/data network to get away from the management of both, we’re certainly not going to separate them again and have a wireless and wired access for clients.

That said, there are opportunities and other areas where there is value, for example location services. Some of the [radio frequency] design. There is going to be a lot of interesting ways people are doing roaming. There will be some interesting technologies from some of these start-ups that will become more widely used.

What is the climate for enterprise IT spending?

If you look at the enterprise, it’s tough to forecast spending. I read all these reports from financial analysts and industry analysts. It’s tough to say what this year’s going to be like, other than to say that there is a lot of activity. But I think buying habits have changed forever. Bandwidth and speeds and feeds – I think that’s old news, like I said. I don’t think you’re going to see people going out saying, ‘I need something faster.’ It’s going to be driven by a change, likely a business change.

What areas in Extreme’s research and development are you excited about?

Looking back over the last couple of years, [ExtremeWare] XOS was more than a three-year effort, so a lot of effort. The new BlackDiamond 10K in December, which uses our 4GNSS technology, that was a three-year investment in ASICs, a very complex system with a lot of carrier-type capabilities. So if you look at all that, those have been big investments. There are follow-on investments, clearly. I think you’ll see more emphasis on software and XOS capabilities. The other area is unified access, which is again mostly software, although we do a fair amount of RF. We’ll look at really filling out the capabilities of XOS and making partnerships there.

As you move more toward software, does that take away from your hardware R&D?

We can do both. In our stackables, we’ve moved to merchant silicon in all of the new products we’ve introduced in the last six to nine months. Because for that type of product, that functionality is good enough. The focus on our ASICs is on the core of the network and the aggregation layer. We’re not trying to build ASICs for all levels of the network. Our focus is on the core. We have a lot more software folks as a percentage today than we did than a few years ago.