Carl Russo: Cisco Avoiding the Core
In a Light Reading interview here at the Supercomm 2001 trade show, Carl Russo, Cisco Systems Inc.'s (Nasdaq: CSCO) group vice president of optical networking, said it may be another 18 months before there are some good opportunities in the core optical networks. He said the disparity between how quickly components are evolving and how slowly service providers are spending on core switches and routers is reason enough for Cisco to keep much of its focus on the metro market, at least for now.
“Go talk to carriers and see what they’re dealing with,” Russo said. “You don’t find them saying: ‘We’re going to double our spending on the core this year!’ ”
Why not? The proof is in the components. “The components those systems are built on are evolving at a pace that they’ve never evolved at before. And the capital spending on [core routing and switching] systems has dried up.
“When you put all those things together and you look at the core, I think it's another 18 months before you start seeing good opportunities in the core.”
It was only two months ago that Cisco decided to suspend its wavelength router, which was built to move massive amounts of data through the cores of service provider networks (see Cisco Kills Monterey Router). The product came to Cisco when Cisco bought Monterey Networks in a stock deal worth about $500 million in August 1999.
Since April, Cisco has focused most of its optical resources toward developing its metro products. This week, in fact, it announced an important upgrade to its ONS 15454 Sonet add/drop multiplexer: the addition of OC192 (10 Gbit/s) interfaces (see Cisco: 10 Gigs and IP Everywhere).
Even the founders of Monterey have reconsidered their original approach to core networks, now opting for a startup product strategy that doesn’t require as much intelligent features in the network’s core as their wavelength router did (see Iris Group Shuffles Executives).
But when would a move back to the core look right for Cisco? It’s hard to say, but there's no doubt that the space is toughening up (see Ciena and Tellium Go Feudin' ). Russo remains convinced that two major parts of core systems would be MEMS (micro-electro-mechanical system) subsystems and tunable lasers. Unfortunately, though, MEMS subsystems aren’t mature enough and there’s no clear-cut tunable laser technology that is worth making a big bet on, he says. "And I do think core networks will be more optical, not all optical.
“So I think core systems are coming along, but the evolution rate is so fast -- just look at Xros (now owned by Nortel). They built a system based on components about 22 months ago and those components are obsolete.
“You want to wait for either some [service provider] dollars to be spent [on core systems] or for a technological slowdown, so that you don’t end up building a system with components that will be obsolete in so many months.”
- Phil Harvey, Senior Editor, Light Reading