Are the parts as great as the whole?

Th Acf13e
Th Acf13e

As some of you may know because it has suddenly started to show up in your already overflowing mail boxes, the crew here at Lightwave recently launched a new publication called Integrated Communications Design. The magazine covers the interplay of fiber optics, electronics, and embedded software in the design and development of next-generation communications systems for carrier and enterprise networks.

I interviewed some of the folks at Avici Systems just before Thanksgiving for an article slated to appear in this month's edition of the new magazine. Much of the interview focused on the company's use--or, rather, its lack of use--of third-party suppliers for components when building its terabit switch/router. It's lonely on the distant edge of electronic technology, Avici discovered. Third-party vendors had little to offer, and the company had to design several application-specific integrated circuits (ASICs) in-house to meet its needs. It turns out that this situation suits Avici just fine. The system's architecture and the development of electronics to support it are Avici's strength, the company spokesmen said.

The lack of available electronics contrasted sharply with the company's optical requirements, however. By the time a carrier is ready for a terabit switch/router, it already has a fiber-based wide area network in place, the Avici spokesmen stated. "The laser technology, the clock and data recovery technology, the test gear sets--everything has already been more or less predeveloped by the physical-layer people for the wide area," reported one of the company's cofounder, Larry Dennison. "There's a fairly long roll-out cycle to deploy fiber in the wide area. By the time someone actually is interested in buying one of our systems, the component people have had a chance to stabilize their offerings for a good year." Thus, Avici has no problem turning to third-party suppliers for optical components.

This conversation came to mind recently when I read an analysis by Ryan Hankin Kent Inc. (RHK) of Cisco Systems' strategy to become a major player in optical networking. Cisco recently filled a hole in its product and technology line with the acquisition of Pirelli's optical systems business (see page 1 of this issue). However, the analysts at RHK point out that Cisco still doesn't have something possessed by the companies it is trying to catch in the optical-networking race: an in-house optical-components business. What's more, the analysis quotes Cisco's director of acquisitions, Amar Hanafi, as saying that the company doesn't necessarily feel the need for an acquisition in this area. Cisco has always bought "best-of-breed" components for its products, Hanafi is reported to have said; why can't it do the same in the optical field?

The RHK report points out that Cisco's competitors-particularly Alcatel, Nortel Networks, and Lucent Technologies-rely on their in-house components divisions to provide a fresh supply of cutting-edge optical technology. I think at some point, Cisco will have to determine where it wants to bring value to the market through its products. Certainly, the new terabit-speed GSR system the company recently unveiled can be seen as a direct analog to Avici's situation. The optical interfaces aren't critical to product differentiation here, and third-party suppliers should be able to provide adequate technology.

However, Cisco's other optical ventures--the new assets from Pirelli, the Wavelength Router from Monterey Networks, as well as the optical systems from Cerent and Pipelinks--may not prove as photonically generic. Certainly, ASIC and architectural expertise plays a role in each of these products, but one can expect that the company will want to incorporate such promising advances as tunable lasers, ultra-long-reach optics, and photonic switches into its offerings to keep pace with its adversaries. As RHK points out, the company is gambling that third-party suppliers such as JDS Uniphase will keep up with the Lucents, Nortels, and Alcatels of the world. If they can't, Cisco will remain one step behind its competitors--or it may be shopping for another acquisition after all.

Stephen M. Hardy
Editorial Director and Associate Publisher
stephenh@pennwell.com

More in Mergers & Acquisitions