The 2002 recession in the U.S. communications infrastructure equipment sales is a difficult period for communications semiconductor vendors. History shows, however, that such times provide the biggest opportunities for market-share gains for the survivors. Where do the gains come from? They don't come from novel protocol opportunities, never-seen-before services, or revolution. Market-share gains come from common-sense execution—that is, execution in bringing ever-increasing value to a customer set, which in turn brings ever-increasing value to existing customers.
Customers, of course, want everything, and not just everything they know they need, but also everything they think they may need. They want everything now—as long as it's free. After all, in these times, everyone is cautious about the future. Decisions to start new projects by equipment companies are driven based on the service providers' requirements for additional equipment. But U.S. carriers aren't demanding new equipment for the purpose of building new services, at least not in 2002. The economics of the time drives product sales based on continued deployment of existing services in a cost-effective way.
That brings us to the third point. Customers want everything for a price that allows them to reduce their costs enough to make their operations director actually smile, even after the marketing director hands over part of the cost savings to the service providers, so that the salesperson still gains the sale above the shouts of desperate competitors.
Let's see how this can be demonstrated in an area of semiconductor communications familiar to us at Wintegra. The access infrastructure is the strata of the network that connects users to core networks. It has not been hit as hard as the optical core, so it becomes a good "middle of the road" example.
What do customers want in this space? Everything, but more specifically, what does "everything" mean here? It turns out that everything is the sum of all access network types: DSL, cable, wireless, optical, and leased-line, put across all network geographies worldwide: U.S., Europe, South America, Asia, etc. Taking the combinations of these technology-geography products reveals some interesting needs. For instance, the leased line, U.S. combination produces a focus on frame-relay (FR) access technologies, while DSL across multiple geographies requires ATM functions. What about the revolutionary all-IP networks heralded in 2000? Well, IP rides easily on top of FR and ATM in the U.S., but more interestingly, in Asia it rides directly over Ethernet metro fiber installations. Is it any wonder then that IP services tend to show up in Asia already? VoIP has a large installation there. IP DSLAMs are increasing in popularity. Korea Telecom recently announced that it would no longer buy ATM DSLAMs in favor of IP DSLAMs.
These cross-product requirements have created an intention by equipment vendors to solve many problems with multiservice boxes. The idea that a vendor can soft provision many types of protocols has been a lure, even if the provisioning isn't passed on to the service provider as a feature. If I can create one card that supports frame relay, ATM, and IP, do I really care if the service provider has the ability to soft configure it? The equipment vendors may simply soft configure the cards while in inventory.
This gives them maximum flexibility to meet their customer demand mix without a plan in advance for how many cards of each type the service providers need, decreasing their lead time, and reducing their cost. They could definitely go further and offer a new soft-provisioned card as well. This, in turn, would decrease service provider costs and improve their lead times, so such a card should command a higher value. This effect could even trickle all the way back to the semiconductor provider.
When do they want this technology? Equipment customer engineers actually want technology early. Their bosses only want to pay for it as late as possible. One lesson from the post-2000 communications recession is that if it doesn't exist yet, it might never exist. This has created a "prove it" mentality among equipment vendors. A request to prove it typically sends the marketers scurrying for cover and the engineers to their schedule spreadsheets. Getting communications chips fully robust is very difficult. If it were easy, everyone would be doing it, but that's not the case. So execution is really a great leveler. Execution makes a pivotal difference in market share.
Finally, what do customers want in the way of cost? Always less. But what is cost? Cost has various forms: dollars, people, and time. Any of these costs can be reduced. Perhaps if a compelling case is made for all three, then skeptical engineering vice presidents of successful companies might believe that cost reductions could be made in one or two of those areas.
The answers are anything but revolutionary: handling multiple services, proving it with execution, and offering several dimensions of cost reduction. Yet while they fly in the face of an all out revolution, they seem to be the new market-share driver—at least for the next year or so.