As embedded systems proliferate, developers need to become even more serious about delivering working software on time. Missing the market window is often fatal for profits and may mortally wound a company, and the inability to cost-effectively design and verify complex SoCs and embedded software is a leading cause of delayed products. According to an Embedded Market Forecaster’s study, an alarming 54% of embedded systems projects are late. Coupling this figure with another research projection – that embedded software soon will account for over 12% of the value of a car – it becomes obvious that pressure for on-time embedded software delivery will only increase.

One of today’s highest profile and highest volume “users” of embedded software, the wireless industry, is characterized by rapid, dramatic technology changes with less than a two-year development cycle and an equivalently short life cycle. The automotive electronics industry is working toward reducing the typical two- to three-year development cycle down to one to two years but with a life cycle of 10 years or more. Although faced with the same electronic development timelines seen in the wireless industry, the automotive industry has no choice but to be even more focused on the reliability aspects of their designs since many of them are targeted toward, or interface with, safety-critical applications.

Design methods have simply failed to keep pace with the growth in complexity and more productive approaches are now needed to address a new set of challenges. In physical and mechanical systems design, model-based design methods are successfully employed to better manage increasing complexity. Model-based design methods are equally appropriate for the design of embedded electronic systems, since modeling allows detailed evaluation of design alternatives well in advance of actual system implementation. Electronic system modeling certainly isn’t new, however new generations of embedded systems in automotive and other areas are made substantially more difficult due to the large number of networked embedded processors that must operate in timing-critical environments.

Achieving the issue-free quality required by the automotive industry, with very high volumes of code and a faster-than-ever development timeline, requires new methods and technologies. The sheer computational requirements demand a breakthrough modeling technology; technology with the flexibility to enable architectural tradeoffs, performance to enable software design, and accuracy to faithfully interact with timing-critical functions such as ignition timing, fuel-air mix, and air bag deployment.

Third-generation electronic system level (ESL) tools and methodologies provide such a breakthrough by improving visibility into the complete system, improving the ability to debug, and improving the ability to control execution. The overall result of is to shorten the design process on average by six months.. Improvements like this allow early market entry, which often enables new products to capture much more market share.

One third-generation ESL method that is providing great advantages to its users is virtual system prototyping (VSP). A virtual system prototype is a software model of a hardware system that runs at speeds near those of the actual hardware and does so with cycle accuracy. This combination of speed and accuracy allows critical control functions to be faithfully replicated in a VSP and explored just as though it were the actual hardware. After the architecture has been optimized using the VSP, it can then stand in for the hardware, enabling a very early start to software development.

Over the past decade, the wireless industry has provided the laboratory in which embedded hardware and software design and test methodologies were created and tested, with high volume deployment of complex systems that were developed to tight timelines. I believe that the automotive industry is a new frontier for embedded systems, and that the techniques and technologies that have been so effective in the wireless industry will go on to serve the critical control functions in the automotive industry. I think the future will see virtual system prototyping becoming the defacto standard for the design and development of critical automotive control functions such as airbag deployment, drive-by-wire and power train.

by Alain Labat, president and CEO, VaST Systems Technology Corporation