As more systems companies move to design their own chips and electronic products, what are the key EDA trends impacting the pace of innovation?
The electronic design automation (EDA) industry is experiencing strong financial growth and is making a vital contribution to the success of much larger semiconductor and electronic systems industries. As more systems companies move to design their own chips and electronic products, what are the key EDA trends impacting the pace of innovation? Find out in this interview with Niels Faché, Vice President and General Manager of PathWave Software Solutions at Keysight Technologies.
Trend 1: Electronic product design is moving toward domain-specific orientation. What impact does domain-specific design have on EDA tool developers and users?
A: It is not enough anymore for product developers to just consider the traditional specifications for a chip or board. They must also now consider the context in which their products will be integrated and used.
Drivers of design for context in product development teams include increased system complexity, more demanding performance and cost requirement tradeoffs, and shorter development lifecycles. To address these issues, EDA vendors and users are seeing closer collaboration in the ecosystem from developers of components (such as an RFIC), to a sub-system (such as a radar), and a system (such as an autonomous drive system) to address integration challenges and optimize performance.
Design-for-context raises several challenges and opportunities for EDA tool providers such as:
Design for context requires closer partnerships between EDA companies with greater interoperability between EDA, computer-aided design (CAD), computer-aided engineering (CAE), and test tools. It also calls for better integration of EDA tools with product lifecycle management (PLM) systems and greater investment in simulation and test process and data management to raise productivity.
Trend 2: Chip usage in all types of products is becoming ubiquitous and the semiconductor industry is now serving an increasing number of customer segments. How is this affecting the EDA industry?
A: Chip demand currently outstrips supply, and this situation clearly has worsened during the pandemic. On a recent trip to Europe, Keysight’s chip design and manufacturing customers confirmed that demand is 30 percent higher than supply. Some chip fabs are fully booked for the next 2 years. However, companies are adding foundry capacity over the next 18 to 24 months that will likely result in a re-balancing of supply and demand.
The semiconductor industry is cyclical and there have always been demand cycles in chip manufacturing that have a downstream effect on the EDA vendor community. For example, automotive has been a cyclical industry for a long time. Other applications such as consumer and healthcare have occupied fab capacity during automotive down cycles. The diversity of applications and industry segments is helping to keep fab utilization high.
Secular growth is strong and the “electrification of everything” is massively increasing the need for new chip sets. Simple 8- or 16-bit microcontrollers are no longer sufficient for many applications that need more advanced computational processing and connectivity. Start-ups continue to sprout up at a rapid pace creating new design starts and innovative products. The fabless model allows the industry to handle a growing spectrum of applications while making efficient use of semiconductor manufacturing capacity.
There will be a continued need for EDA products to meet new design functionality and validation efforts. Design teams need better tools, IP blocks, and consulting services from EDA companies. Expanded use of chips in customer markets is a very positive development for EDA vendors whose growth and success rides in part on design starts and wins. More designs started means more demand for engineers and the EDA tools they use to get their jobs done faster with intelligent automation and higher productivity.
Trend 3: Customers are demanding that chips and electronic systems last longer and function correctly throughout their lifetimes. This is particularly important in safety-critical markets like automotive and mission-critical markets like data centers. How do EDA tools address product aging, quality, and reliability concerns?
A: Design for reliability is nothing new for Keysight because the company’s instrument products have very stringent lifetime requirements. Reliability becomes a net positive only if fully embraced throughout design, manufacturing, and test processes. Keysight incorporates reliability best practices into its instrument product life cycles that have positively influenced its design and simulation tools over many years of development. Keysight’s own internal tool users and commercial customers are increasingly interested in making sure circuits stay within electrical and thermal limits. This seems simple, but can be quite challenging, especially when considering environment and process variation.
Chips are increasingly part of a much larger system interconnected through different packaging technologies. Modeling these interconnect and packaging details is what makes designing for reliability so difficult. For example, space applications require built-in redundancies and special design patterns to increase radiation hardness. This approach is also used in other mission critical applications such as healthcare. Reliability and aging (wear out) requirements are gaining importance in IOT, automotive, and consumer products where once aerospace and defense was the leading driver.
As the cost of mistakes go up in these new applications, the importance of simulation on design quality increases. Keysight’s PathWave design tools enable simulation and analysis of signal and power integrity and electromagnetic effects, which directly affect quality and reliability. The impact for EDA tools could become even more significant if design customers push for simulation-as-signoff from the industry. This requires validation of software through independent test suites or test bodies. EDA tools and IP are also helping to predict and avoid field failures. Real-time data collection and analysis using embedded sensors and AI/ML software techniques promises to address silicon product reliability and aging soon.