The Light Reading Interview

In The Spotlight: Jack McMaster

KPNQwest NV http://www.kpnqwest.com is one of Europe's new-style service providers. It's rolling out a fiber backbone connecting 39 cities in 13 countries, which is linked to the U.S. backbone operated by Qwest Communications International Inc., http://www.qwest.com. It's also building a string of colocation centers. Qwest offers a full range of wholesale and retail IP services.

The company was founded by Qwest and KPN Telecom BV http://www.kpn.com, a subsidiary of the Netherlands' incumbent carrier. Last November, it staged an IPO and its shares are now traded under the ticker symbol of KPIQ on Nasdaq and the Amsterdam stock exchange.

KPNQuest's backbone is based on SDH equipment from Alcatel SA http://www.alcatel.com. Just recently, it picked Juniper Networks, Inc. http://www.juniper.net in preference to Cisco Systems Inc. http://www.cisco.com as its backbone router supplier.

For more background see http://www.kpnqwest.com/pressroom/press_factsheets.asp?choice=factsheetsJack McMaster took over the reigns as president and CEO of KPNQwest in May 1999, having previously worked for AT&T and Qwest.

In this interview, McMaster talks to Light Reading about:

* Why KPNQwest picked Juniper and not Cisco for its backbone routers
* The dangers carriers face when accepting fancy financing deals from vendors.
* Who's winning the race to wire up Europe.
* Why providing bandwidth won't become a low margin commodity business.

Light Reading: There must be more than a dozen operators building fiber backbones in Europe these days. What makes KPNQwest different?

McMaster: We're not just talking about rolling out a backbone. We're actually doing it, and we're already offering services.

Some of the other operators make a lot of noise but haven't even started. They also don't own their own fiber. They're buying bandwidth off each other.

Light Reading: I'm sure some other operators would say they're further advanced than KPNQwest. For a kickoff, GTS - or what used to be called Hermes - started rolling out its backbone several years ago and has much greater reach.

McMaster: Yes, but GTS seriously underestimated bandwidth demand. It only laid four strands of fiber on most routes, reckoning that developments in DWDM (dense wave division multiplexing) would mean that it would always have enough capacity.

That's proving to be a big mistake. We're installing cables with 96 and 122 fibers in them. In fact, the one under the Channel as 144 fibers.

Owning your own fiber, and having lots of it, is crucial to our success. It enables us to turn on vasts amounts of bandwidth to meet demand. It also enables us to control the quality of the services we provide - and that's key.

Light Reading: I've heard the same point made by other carriers, though. And at least one of them - MCI Worldcom - started building its backbone long before KPNQwest existed.

McMaster: I count MCI Worldcom as serious competition, along with Global Crossing - although I think we're ahead of Global Crossing.

Right now, I don't think it's clear who's going to end up being the market leader. But it's like the Olympics. You can already identify the guys that aren't going to win medals - and I would say they include GTS, Iaxis and Viatel.

Light Reading: Doesn't your strategy of owning fiber have a flaw in it, in that KPNQwest has put all of its efforts into building an international backbone and doesn't own fiber within cities - where you need to connect customers?

McMaster: Right now, I can use fiber from the growing number of metropolitan operators in Europe. But at the end of the day, I will need my own fiber in cities.

Light Reading: I guess building all of these colocation centers is an attempt to get customers - such as ASPs (application service providers) and Web hosting companies - to connect direct to your backbone. Is that right??McMaster: Yes. To me, the colo centers are the peanuts on the bar. They're a way of attracting customers for the IP bandwidth that I'm selling on my backbone.

Light Reading: But isn't bandwidth going to become a commodity and isn't providing it going to become a low margin business?

McMaster: I'm not sure how much of a commodity it will become, or how low margin it will be. I see demand exceeding supply for the foreseeable future.

What we're going through now is like the beginning of the petrol economy. People are having trouble imagining how much bandwidth will be required, because it's on a totally different scale to what's gone before it.

Light Reading: Let's talk about how you picked vendors for the key equipment in your backbone. For a start, why did you go with Alcatel for your optical gear?

McMaster: Alcatel's technology is the best, but it's horrible at marketing. (Sighs). It's one of the crosses I have to bear.

Alcatel was the only company that was able to deliver 10 gigabit transmission with 2.5 gigabit connections for routers, in a seamless, efficient way, in a single cabinet in the CO. The other vendors needed three cabinets to do the same thing.

The other vendors tried to say that I didn't need 10 gigabit and that I'd be paying a premium to get it. But that's because they couldn't match Alcatel at 10 gigabits - and I want to maximimise the amount of bandwidth I get out of my backbone. I didn't want to be stuck with a 2.5 gigabit network.

Luckily, I was in a position where I could stick out for what I wanted.

Light Reading: What do you mean?

McMaster: We've raised more than $1 billion through our IPO and by issuing high-yield bonds. As a result, we're not beholden to vendor financing. We've got a free hand to pick vendors on the basis of their technology.

Light Reading: Is that unusual?

McMaster: Yes. Vendors end up offering roughly the same technology, sooner or later. So technical evaluations often come down to arguments over how many angels can dance on the head of a pin. At that point, vendors try and sway the argument their way by coming up with an attractive financing package - and that often wins the day.

Light Reading: I suppose the money you've raised through your IPO and bonds gave you the necessary independence to pick Juniper's new M160 rather than Cisco's GSR12000 for your backbone router?

McMaster:Yes. Juniper's router was a lot faster in the tests we conducted. When we told Cisco the news, they took it so badly. They were really stunned. So upset.

Of course, it isn't the end of the story. We're just like every other carrier. We always have another vendor in the warm-up booth.

by Peter Heywood, international editor, Light Reading http://www.lightreading.com