The conversion to merchant silicon for high-end Ethernet switching is inevitable. As the market matures and hardware features become standardized, there is less need for systems vendors to develop specialized ASICs to differentiate themselves. ASIC vendors and their users will move to other functions, like security, where they can differentiate and add value. The economics simply don't work out.

Service providers today, and their equipment providers, are preparing for the next generation of mixed services. Those include voice/voice-over-Internet Protocol, IP virtual private networks, Ethernet LAN extensions, broadband access and IP storage. Service providers need to accomplish this quickly, while continuing to deliver today's traditional revenue-generating services. These services pull system requirements in a variety of directions, resulting in more complexity. With market windows narrowing and the unfortunate corollary to Moore's Law that chips double in development cost every 18 months, something's gotta give.

How can the industry make the transition in a painless, open and flexible way that creates a win-win for both chip and systems companies? The best way is to consider architectural flexibility so that the applications driving networks are considered at the design level. While this makes the development of high-end Ethernet switching more complex, it also drives innovation and opens an opportunity for these organizations to lower their costs and speed their time-to-market.

Despite the cultural bias for internal silicon, the growth of merchant silicon is inevitable. Many different technologies have already gone that route: Microprocessors, cell phones and low-end networking technologies are all merchant- based silicon. As evident through these examples, the only true architecture that succeeds is one that's flexible and able to fit a variety of solutions. The rapid rise of Intel is evidence of this. In the end, technologies that cut costs and allow companies the ability to reach the market more quickly will prevail. This is where merchant silicon has succeeded and ASICs have so often failed.

Ethernet systems are now being differentiated on the basis of software, service and brand. High-end systems certainly have hooks in their ASICs to deliver performance. However, a flexible merchant solution can offer the same advantage, with lower development costs. To remain competitive, OEMs need to work with silicon that is cutting-edge in functionality and performance. IC suppliers, in turn, need to be on top of advances in silicon design and manufacturing technologies as well as up-to-date on communications standards and evolving market requirements.

With acceptance of merchant silicon in the high-end Ethernet switching market, chip and systems developers can focus on engineering networks and outsource this to their customers, who can in turn focus on running their businesses instead of their networks. This is pure economics — specialization increasing efficiency.

1323myfceostaccato

ULTRAWIDEBAND RISE

G. Roberto Aiello, President and Chief Executive Officer, Staccato Communications Inc., San Diego

The communications sector has been peppered with pessimism over the past few years, which makes it exciting and invigorating to experience firsthand the innovation, support and enthusiasm that has surrounded the more-recent emergence of ultrawideband (UWB) technology.

The rising tide in this wireless technology is evidenced by the fact that during the first quarter of this year alone, three UWB startups together have attracted more than $100 million in new venture-capital funding. There are a few more tangible and telltale pieces of evidence that our industry is on the rise again than to see the resumption of VC announcements and IPOs making headlines.

The forecast became sunny for UWB in February 2002 when the Federal Communications Commission allocated the largest unlicensed block of spectrum ever for consumer use. The move represented a huge vote of confidence for a technology that promises to replace wirelessly unwieldy and bothersome cables while delivering up to 480 Mbits/second of bandwidth for an ever-increasing number of Universal Serial Bus and Firewire applications that span the consumer electronics and personal computing markets.

As it takes a village to raise a child, similarly it requires the embrace and empowerment of many companies and industry organizations to deliver and deploy a successful technology. UWB is fortunate to be nurtured by a growing ecosystem including the Multiband OFDM Alliance, WiMedia Alliance, 1394 Trade Association and Wireless USB Promoters Group as well as many of the most influential players in personal computing, consumer electronics, semiconductors and mobile communications.

Even with such staunch backing, UWB has relied on a series of technical innovations to deliver on its promise. Just as ultrawideband evolved from the early single-band impulse radios to a more fully functioned version based on multiband OFDM, our next hurdle is to deliver the UWB industry's first all-CMOS radio. The advantages of CMOS already are legion in the communications world. CMOS represents UWB's best opportunity to take advantage of Moore's Law while meeting all the regulations set down by the FCC regarding low power consumption and price with high performance.

The first implementations of high-speed RF designs in mainstream digital-CMOS process technology only debuted in the mid-1990s. To date, only a select few RF design teams worldwide have been able to deliver on the promise of CMOS. Compounding the effort, no company has yet implemented a high-bit-rate, all-CMOS design.

At Staccato Communications, our belief from the start is that a technology innovation such as digital CMOS integration is the only economical way to deliver cost-effective, highly functional, interoperable UWB-enabled products. Once this innovation is realized, as with Bluetooth and 802.11 before it, ultrawideband will begin its opportunity to become just as pervasive.

Vince Graziani, Chief Executive Officer