>Scotland's Project Alba is an ambitious attempt to create a "design factory" for system-on-a-chip design
outside California's Silicon Valley. Can such a center succeed in Scotland and elsewhere in the world?
Cadence Design Systems thinks so and plans to aggressively pursue its design factory concept globally.
However, it remains to be seen whether the technical and business infrastructure that has tied so much IP
(intellectual-property) development and integration to California can be replicated elsewhere. Retooling
universities, overcoming legal obstacles and fostering an entrepreneurial spirit are just some of the challenges
facing those who want to take system-on-a-chip design to new places. The idea of a design factory, said Bob
Leach, senior vice president of market development at Cadence, is a "repeatable, engineered process" for
some particular aspect of chip design. Cadence's Livingston, Scotland, design center is intended to meet this
definition by focusing on IP integration in targeted areas such as mobile digital communications, digital imaging
and digital TV. But Cadence has more than Scotland on its mind. "We have created a repeatable process for
helping sovereign governments first assess the unique capabilities in their region, then target a growth market
segment and create a plan to become a highly attractive location for design work in that particular market
segment," he said. Cadence is believed to have approached two U.S. states, as well as several foreign
governments, with Alba-like plans. Leach declined to specify where Cadence is talking, but he pointed to Brazil,
Israel and Australia as promising locations for system-on-chip design and IP development. Commercial IP
development today is heavily centered in Silicon Valley. According to data from HTE Research, about 64
percent of the $390 million commercial IP market in 1997 originated in California. IP integration is arguably
more dispersed, with companies such as Alcatel, Siemens in Europe and NEC in Japan running very active
system-on-chip design operations.

Steve Szirom, president of HTE Research, is skeptical about Cadence's concept of exporting system-on-chip
design factories. "I don't think it's a workable idea," he said. "If you're talking about creating a skill center like
Silicon Valley, in most areas of the world there are two things missing-entrepreneurial fervor and a
semiconductor company infrastructure from which talent can be drawn." There are significant reasons why so
much IP development takes place in Silicon Valley, said Howard Sachs, chairman of the Virtual Socket
Interface (VSI) Alliance. "There's an excitement and a business environment here and a tre-mendous amount of
knowledge, and it feeds on itself. Also, there's a lot of money to start companies. If you want to make a lot of
money you come to Silicon Valley. The infrastructure is here."

But others say that IP development and system-on-chip design can and do occur in many places. Jim Tully,
senior industry analyst at Dataquest Inc., noted that a great deal of IP development takes place in the U.K., not
only for European companies such as ARM, but also for U.S. companies such as MIPS, DSP Group and
Mentor's Inventra division.

In fact, said Tully, two-thirds of the world's design houses are located in Europe and many have developed their
own IP. "I don't think there is anything that ties IP development to Silicon Valley," Tully said. "It's the marketing
that happens in the United States."

"In terms of the ability to attract talent and do the work, there are many cases where being outside the U.S. has
some advantages," said Reynette Au, vice pres- ident of marketing at ARM Ltd. (Cambridge, England). "The
talent pool tends to be focused on a small number of companies, so you can minimize the job hopping that
occurs."

Au noted, however, that most of ARM's customers are in Silicon Valley. Thus, it has been important for ARM
and for other European IP providers such as Sican and Integrated Silicon Systems to have Silicon Valley
offices. Opportunities aboundFrom Cadence's point of view, there are several locations around the world just
waiting to become system-on-chip development centers. One place with potential, said Leach, is Brazil.
"People in industry and trade are telling me they're turning out 2,000 PhDs in EE every year, published in
English, and 60 percent of them can't find work," he said.

Australia, said Leach, produced three of the top 10 new products at Comdex last year and could be a major
source of IP in areas such as multimedia and tele- communications. "They have the broadest implementation of
broadband, spe- cifically XDSL technology, anywhere in the world," he said. Israel has produced a lot of
innovative IP, said Leach, but many of the developers there lack access to international marketing networks.
There are also places that don't look so good. "There are jurisdictions in the world that are terrible places to
consider commercial exchange of IP, because IP rights are so weak," Leach said. "At the top of that list is
China and there are questions about India, Korea and Taiwan." Szirom said that India has become a hot spot
for software development and that some IC design activity has started there. "There is some IP, but not too
much," he said. "It's mostly contracts and design support." One of the best-known Indian IC design houses,
Arcus Technology (Bangalore), was recently acquired by Avant!'s Galax! subsidiary. Arcus has an extensive
portfolio of digital and mixed-signal IP, including Firewire, PCI and USB cores; a 32-bit embedded processor,
and various communications and networking cores. Since large Japanese semiconductor companies are
doing IP integration work, it's perhaps surprising that there is almost no third-party IP activity in Japan. The
problem there, said both Leach and Szirom, is largely a cultural matter-engineers don't want to leave the
security of large companies to start their own. One of the few Japanese third-party IP vendors is Nomura
Semiconductor Technology (Kyoto). Making it work If someone wants to go to some previously untapped region
of the world and set up a system-on-chip design factory, there are several important considerations. First, said
Leach, there must be a high-quality university structure that does innovative chip-design research and turns out
a sufficient number of engineers. Some retooling is probably necessary in the university environment. Many
universities are still teaching transistor-level or gate-level design, as opposed to system-level design. For this
reason, Cadence is working with several universities in Scotland to put together a comprehensive educational
program for system-on-chip design.

The interaction between universities and government is important. In Japan, Leach noted, private industry can't
give universities money for research that might have applicability for the donor. "This makes industry-university
collaboration difficult and has made engineering education almost irrelevant," said Leach. Secondly, there
needs to be a cooperative government. One politically challenging task, said Leach, is to identify an "area of
focus." If a given country decides to focus on telecommunications, for example, that could annoy business
interests that are pushing for digital TVs. The legal structure is also a big consideration. Nations differ widely in
how strongly they protect IP, as well as the kinds of obstacles they throw around IP commerce. Brazil is a case
in point-in the pharmaceutical industry, Leach said, Brazil was once forcing people with patents to divulge them
to others. In the U.S., Leach said, the legal system provides good IP protection, but the system is not efficient.
"There's a pretty ugly set of existing case law that contract lawyers have to work their way through," he noted.
"Each transaction becomes custom. It can take nine to 12 months to negotiate the transfer of IP for a product
that has a usable shelf life of nine months." For these reasons, Cadence is working with Scotland to create a
system of "private law" modeled after what was done in the international maritime industry. It remains to be
seen how well Project Alba succeeds. HTE Research's Szirom said he thinks it will have recruitment problems.
"Most of the engineers and designers I know don't want to go to Scotland," he said. For more technology news,
visit http://www.eet.com