The decline in ASIC and ASSP design starts over the past decade and the increasing difficulty, cost and risk of hardware design, have all been well documented.

However, at the same time the number and diversity of silicon-powered systems has exploded. Think about everything on the market today – numerous cellular phone models, mp3 players, satellite radio receivers, wireless headsets, digital cameras, plasma and LCD displays, DVRs, WLAN access points, GPS navigation units, Bluetooth devices, …

In short, companies are designing more electronic systems, yet fewer ICs are being designed. What does this mean? It means that consumers of semiconductor devices are adding more value beyond the chip. But does this also portend gloom and doom for EDA?

Hardly.

EDA stands for Electronic Design Automation, not Semiconductor Design Automation or Integrated Circuit Design Automation. The key word here is "Electronic". So. this is a call to the EDA industry: Look beyond the chip! Help electronic system designers design better products!

How to do that, you ask? There are many ways, but one way is to help designers cope with the increasing amount of embedded software in products today.

To design digital logic circuits, the EDA industry created a plethora of fast circuit simulators, synthesis engines to generate gates from multiple input languages, and tools that assist the verification process at various levels of abstraction. Current and future designers need similar levels of automation to efficiently create their software-dominated systems.

They need fast simulation of algorithms and multi-processor systems. They need synthesis from higher-level languages to efficient firmware implementations (whether it’s code for a DSP, an FPGA bit stream, or microcode for a controller). They need help verifying the final system implementation matches the original specification.

Broadening the EDA industry to better help system designers won’t be easy. There’s technical risk – the tools are difficult to build. And not only is additional expertise needed here, but product requirements are fuzzy and existing system design flows are diverse.

On top of that, new business models are needed. These products will no longer sell for hundreds of thousands of dollars per seat, so they will need to be sold differently. However, the products will be able to target a much larger user base – there are easily over 1 million embedded software programmers.

Given these challenges, why broaden? Why not continue developing and selling traditional EDA products, like IC place and route tools? The customers and business model are well known. And while the customer base is slowly shrinking, it’s still large enough for a few players to make a living. However, each new process node presents a taller hurdle for traditional EDA customers to jump over.

By contrast, system designers and developers are attempting to build many more new gadgets and devices. They are putting together various combinations of semiconductors, software and peripherals in hopes of filling an end-market need. There is little automation in their current design flow. Yet these folks need help - and they are willing to pay for it. This new market is both huge and untapped.

Traditional EDA won’t go away anytime soon. There will always be a set of applications that will justify the cost and risk of developing an ASIC. The number of such applications is dwindling, further concentrating the traditional EDA customer base.

If you’re searching for how to grow the EDA industry, my advice is to look beyond the chip.

*****************************