2005 Emerging Trends: A final word on the topic(s) ...
We ushered in 2005 with a 2-part discussion about the EDA industry and several closely aligned disciplines. In Part 1, we talked about EDA and the Embedded Software industry (ESW). In Part 2, we talked about EDA and the Electronic System Level design industry (ESL). This week, we conclude the discussion with an addendum. As of January 10th, Alain Labat has been officially named as CEO at VaST Systems Technology. In advance of the announcement, he sent the comments below which address the questions posed in Part 1 of the 2005 Emerging Trends articles. I am grateful for his feedback and wish him well in his new role. Labat's comments follow on those in the previous two articles from:* David Stewart at Critical Blue Meanwhile, as the industry looks to DesignCon, where we'll hear from the CEO's of the Big 3 EDA players Wally Rhines from Mentor Graphics, Mike Fister from Cadence, and Aart de Geus from Synopsys it will be interesting to note how these folks position their organizations with respect to both the world of embedded software and the world of ESL. ************************** 1) Why isn't there a closer alignment between the EDA and the ESW industries? Alain Labat: Until the last few years, hardware design and software development were very different disciplines, not particularly dependent upon one another and performed sequentially. Software development began when the hardware was nearly completed. The world of system on a chip, with growing amounts of embedded software, has begun to change things...but old design habits die hard. 2) Do you see a future synergy between Embedded and EDA? Alain Labat: If EDA is to grow and prosper, it will be forced to supply solutions to the new world of embedded software. Gartner Dataquest and other industry trackers suggest that up to 70% of all chips designed today contain embedded processors and that over time the most cost - effective way to design functionality will be with embedded software. Ideally, the embedded software development needs to begin concurrently with hardware design and both need to work from a precise architecture model. 3) If so, what do you foresee this looking like? If not, why? Alain Labat: It is logical to assert that EDA will wake up and recognize the growing need for stand-alone (not shipped by the RTOS vendor) EDA tools to serve the embedded market, but logic does not always prevail and does not seem to be making very great progress in this case. At VaST, we have begun to think that a new category will arise separate from either EDA or embedded. Here we call that new category "architecture-driven design". The technical prerequisite for this method is model-based development using very advanced modeling technology coupled with a whole new way to simulate. There are already a number of solutions in this category. In fact, the last few years have seen production-worthy solutions that support the development of a software virtual prototype of the system that is both timing accurate and can run at real hardware speeds. This type of software virtual prototype makes it both possible and practical to develop and verify the hardware and software concurrently. 4) The EDA industry is looking for a way to increase their market. Is involvement through M&A with Embedded system tool enterprises one-way to accomplish that? Equally, could embedded software tool vendors broaden the market by acquiring EDA companies? Alain Labat: Yes, mergers and acquisitions are certainly possible, although as alluded to previously, the EDA industry seems to still see itself as a supplier of tools for traditional hardware developers, NOT for system designers and certainly NOT for software developers. An interesting analogy is how the U.S. railroad industry missed out on the transport of goods in the late 19th century. It missed, NOT because it could not provide the technology required, but because it could not conceive of itself outside its traditional "move the people" mission. To capture the transport business the railroads would have had to change the gauge of the tracks to facilitate the long hall of goods. They did not move soon enough and so the automotive industry came along with trucks and stole the entire goods transportation market... and STILL dominates it 5) Can you envision a unified environment? Where are the main/key bridges that need to be built to two industries/sectors together efficiently? Alain Labat: The Fall Processor Forum was interesting this year. Many of the presentations from the processor companies emphasized that, in the future, it will be the "system" that will be pushed for performance more then the processor. The handwriting is on the wall for the EDA industry. Customers need to design and verify at the SYSTEM level. This movement actually began some years ago with a few startups providing the first hardware/software co-verification solutions. As welcome as these efforts are they fall far short of what industry needs. As systems move increasingly to multi-processor designs, the overall complexity of the system requires that it be architected top down for performance, timing accuracy, power analysis and a host of other attributes and verified in a formal way at each step. We believe that these requirements dictate that the system be represented by a model created in software that will run at hardware speeds with no sacrifice of accuracy. 6) Ultimately, it's a question of the technology. Why don't EDA vendors provide embedded SW development solutions and vice versa? Alain Labat: See my "trains" analogy earlier. Also it is a very, very hard technical problem to solve in an elegant, and cost-effective way. 7) What is the customer's current pain point that is being left unsolved because EDA and Embedded system tool vendors do not provide a unified solution?' In other words, what business opportunities are being overlooked now and in the future? Alain Labat: Complex systems are increasingly loaded with embedded software. Despite the best efforts of architects, many of the architectural problems will actually be discovered by the software engineers (or, horror of horrors, by the customers). Only when the real software is written, is it possible to definitively assess the system performance. It is then that you find bus-contention problems, cache issues that degrade the overall system performance, excessive power consumption, and inadequate bandwidth to transfer the data load. The software MUST be developed concurrently with the hardware to solve these issues.
Peggy Aycinena owns and operates EDA Confidential. She can be reached at peggy@aycinena.com
|