No one will argue with the nuclear-winter description that has been applied to the optical-networking downturn. Right now, I think anything that will improve the carrier's cost structure is important. The success of multiservice provisioning platforms, coarse wavelength division multiplexing (CWDM), and enhanced protocols like the generic framing protocol and multiprotocol label switching (MPLS) will give carriers more value for less dollars and improve the efficiency of their existing legacy Sonet infrastructure. Although these are incremental rather than revolutionary improvements to the network, they are critical if carriers are to move toward making money on data. From the point of view of IC suppliers like Analog Devices, the drivers are the same: Make the chips faster, more integrated, and smaller in size, and reduce the power consumption. And while you're at it, reduce the price.
Recently, I have seen some encouraging signs. There has been a much-needed consolidation at the optical components level and even some signs of pullback in the IC sector. Consolidation is required for a healthy recovery, otherwise too many companies will be chasing too few sockets when business comes back. Analog Devices' broad base of business means that we can ride out weakness in a particular market like fiber optics and maintain a long-term focus.
Even to those of us who are strong enough to weather this storm, it's still a confusing picture out there. The slump has been longer and deeper than anyone predicted. I still see most of the attention re-focused on legacy Sonet applications in the telecom space. I can also imagine that the metro will feel capacity pressure soon. That's because the carriers are running the networks at historically high utilization rates, which will force them to purchase new metro telecommunications equipment.
With inventories burning off and network utilization rates continuing to rise, I believe there's a general sense that carriers will begin purchasing new equipment late in 2003.