Enterprises are nowdriving the bus
One of the more popular finger-pointing exercises in the aftermath of the telecommunications downturn involved taking analysts to task for their inflated forecast numbers. Many of these abused individuals replied that they were just repeating what they had been told by their vendor sources. And the vendors said they were only reporting what their customers were telling them. Of course, who the customer is depends on where you are in the food chain, so the blame kept getting passed upward toward the top of the pyramid. Component vendors said their numbers came from systems vendors, systems vendors shrugged and said they were only reacting to what the carriers told them, and the carriers-well eventually someone decided to blame some numbers cruncher at UUNET. We'll all do better, everyone said, and market research firms began to tout methodologies that involved discussions with both suppliers and carriers.
Of course, this all left out the final component in the market equation: the users of communications services who represent theultimate customer for everyone involved. If the average resident or business user isn't happy, that displeasure reverberates to the wafer level. And it appears that this fact has finally hit home, because right now the service user-particularly in the enterprise-is calling the shots.
How do we know? Remember how IP was going to take over the network from end to end? Well, business customers prefer the Ethernet they've come to know and love, and now carriers can't put Ethernet services into the field quickly enough. True, IP/MPLS has established itself in the core and has begun to stretch toward the edge, but when the service hits the customer premises, it had better look like Ethernet.
At the same time, enterprise economics have replaced carrier economics when it comes to components, subsystems, and systems development. In the optical space, transceiver vendors in particular have felt this impact. The work on the 10GBase-LRM PMD is a direct result of cost concerns. So too is the suddenly slower drive toward 10 Gbits/sec in the Fibre Channel community. Networkmanagers don't want to pay for 10 Gbits/sec if they don't need it-and so we have 2-, 4-, and eventually 8-Gbit/sec Fibre Channel transceivers in pluggable form factors that enable network designers to pay for what they want and only when they want it.
But when they want high bandwidth, they'd better receive it, or they'll take matters into their own hands. That's why CWDM companies and free-space optics vendors have been able to supplement their carrier sales by marketing directly to enterprises. Major private firms, some driven by security concerns and government mandates, have developed significant high-bandwidth requirements. Medical facilities intent on marketing their services want to take advantage of the latest imaging technology as well as the ability to collaborate or learn via video. If carriers can't meet those needs, then major end users have little choice other than to take their communications matters into their own hands. And it's now easier than ever to do that,especially in areas where fallow optical infrastructure remains available.
In short, the dynamic is moving from one where enterprises had to just accept what was offered to them to a scenario where they can choose among options, particularly if they have the courage toincorporate optical media into their private networks. And because many optical communications equipment vendors must expand their customer base beyond the traditional carriers, optical technology will continue to become cheaper and easier to use to penetrate this market.
Of course, carriers also like cheaper and easier, so you can expect them to ask for the same benefits that enterprise customers will extract from technology suppliers. Therefore, the new emphasis on the enterprise will only accelerate the trend away from technology-driven product offerings to those based on applications.
Margins will likely continue to tighten as well, at least until manufacturing efficiency catches up. But the growth of the optical end user offers the opportunity for sales volumes that will require dusting off a few mothballed production lines. And it's about time for that.