Published in June 2008 issue of Chip Design Magazine
ASIC design starts continue to fall! Fewer investors support EDA tool start-up companies! Consolidation among fabs, chip houses and software vendors is running rampant! Investment firms – like Merrill Lynch – no longer even consider EDA as a separate growth market! It’s the end of the world as we know it! … or is it?
Few doubt that the semiconductor electronics industry is metamorphosing into something different. All of the major trade journals – those of us that still survive, at any rate – spend considerable energy over the last few years covering the important trends (see references).
What will be the topology of the new semiconductor electronic landscape? That’s the question that every serious CEO, investor, marketing/PR and editor asked themselves in deadly earnest. It may be that the EDA industry will become part of the bigger picture,
namely the chip-package-board design and manufacturing world. But this will mean a change in the supporting technologies and business models for not only EDA but the other related industries,
such as the semiconductor infrastructure companies. Is it possible to predict what the new world of EDA and semiconductor development will look like? Perhaps – but I’ll get to that shortly.
The current turmoil in the semiconductor electronics world is but an artifact of a bigger change, one that any systems level engineer will quickly recognize. This bigger change is the blurring of boundaries
between all disciplines. In the EDA and semiconductor world, that change manifests itself most noticeably between the increased integration of cross-discipline engineering, such as hardware-software
and analog-digital development. But this need to cross traditional thought boundaries or established disciplines is occurring at a much high level.
Perhaps no where is this more graphically illustrated than in a recent presentation by James Burke’s and his “Knowledge Web” presentation (www.k-web.org). Most of us remember James Burke from his awarding winning and just plain fun TV series of the 1970’s called “Connections.” James has been busy since then, focusing on the role of interconnections and relationships in innovation. His current project – called the Knowledge or “K” web – demonstrates how all knowledge is interlinked. This is important because a solution in one knowledge domain, say embedded board level technology, might be very useful to another domain, say EDA technology – and vice versa.
It has been said that change comes from the no-man’s land between disciplines. James notes that today such a no-man’s land might be between nano and biological-information. I would add
that, for our world of semiconductor electronics, there is a no-man’s land is between chip-package and board and product. From my own experience: semiconductor folks talk about ESL. Board level folks
talk about system design. Software engineers talk about systems engineering (CMMI, etc), while network guys talk about systems engineering. We’re all saying the same think – we are, truly. The first
week of my university graduate course I teach on hardware-software development tries to explain these different applications of “basic systems” thinking. The tragedy is that no group wants to talk to the other group. But they have to if they want to meet time-to-market and the demands of complexity.
In fact, several well known companies from our small world of electronics and semiconductor technology are helping on this KWeb project, including Sun Microsystems and Applied Materials.
This may be something the other players should check out, as well. It might be profitable to have everyone interconnected in terms of innovation and development processes, instead of constantly reinventing the wheel.