In the semiconductor design industry there exist three industry mega trends that will, to a great extent, determine who will emerge as “winners” in the future. These mega trends are: the world of “AND” in innovation and execution; the move from System-on-a-Chip to “Systems”-on-a-Chip; and leveraging partnerships. These will impact the skills requirements of professionals in the VLSI world.
World of “AND” in innovation and executionGone are the days when one could have the luxury of focusing on one dimension of innovation and execution. The parameters have been redefined. What used to be referred to as “solid execution” till a few years ago may not be good enough anymore; and what used to be a good example of innovation may not be true in a short time to come.
Today, innovation and execution have multiple facets, and one needs to keep pushing these two envelops.
The first aspect of innovation and execution is product quality. One used to think that designs, sampled to customers in first pass, was proof of good quality. This is no longer the case. The word “quality” has given way to the word “robustness”.
A “functional” device, a good design or software demo to customers are only first steps. Against the new bar of excellence, one would ask, “how many revisions did the design take to get to high volume production? How long and how many passes did it take before it reached the desired levels of yield? Did it meet all the system interoperability requirements?”
Robustness also includes “reliability”. Applications in automotive, for example, demand the highest levels of product reliability and lowest DPPM. This high bar completely redefines excellence in design robustness. Similarly, increasing number of applications today stretch the quality and reliability of software. Developing software to demonstrate prototypes to customers is no longer adequate. The goal is to develop truly robust software, production-quality software that can be shipped in high volumes.
It is no surprise, therefore, that the acceptance criterion in terms of product quality—the most fundamental requirement ofall—has undergone a dramatic transformation. This requires an intense focus on “cost of quality”. Every VLSI design engineer, every software developer in the embedded world, needs to reflect on this and must set a higher bar of excellence.
The second aspect of innovation and execution is cycle time and efficiency. The increased complexity of the products has put enormous strain on this dimension. Customer expectations and time to market have risen, and rightly so. Yet, the scope and complexity of products have changed by orders of magnitude. How does one go up the complexity ladder in shorter time? One option is to use more people. However, this option does not scale due to cost considerations and practical limits of availability.
What is needed is really increased efficiency—scaling the complexity in shorter time cycles, with fewer resources! This requires a combination of deep domain know-how and high focus on automation. Yet again, this redefines the bar of excellence and success completely.
The final aspect of innovation and execution is delivering leadership and differentiated products to the customer. Developing a leadership product is no longer a one-dimensional problem. A product that innovates and achieves the highest speed but
is not cost competitive or drains a lot of power or has a poor signal-to-noise ratio, does not make the grade. True, the expectation bar in each of these dimensions varies with customer application.
However, what is common is the macro change: customers want the highest performance, they want the lowest price, they want the lowest power devices, and they want best linearity. Looking at it from the customer’s angle, it is of no surprise. Their world has become intensely competitive. To lead therefore, they need the highest levels of cost-performance from the semiconductor world.
Therefore the first mega trend is the new world of “AND” that we are entering and how this is redefining execution. We need product quality AND cycle time/efficiency AND deliver leadership products.
Product quality itself is getting redefined to include the highest levels of robustness. Cycle time/efficiency expectations are getting redefined as the baseline complexity is moving the other direction.
Leadership products are getting redefined and becoming a multi-dimensional variable, calling for optimization along different vectors simultaneously.
Move from System-on-a-Chip to “Systems”-on-a-ChipEach building block in an SoC is not a component today—it is a subsystem and, in many cases, it is even a complex system. The complexity this poses to the systems design domain, to system validation, to the SoC integration world, to design-for-test or physical design is immense.
Part of this migration includes an entirely different level of focus that leaders must have in areas such as systems and software, analog and RF. The move to Systems-on-a-Chip naturally implies a much deeper level of systems know-how. Performance modeling and tradeoffs, as an example, once done between sub-systems will now have to be done across systems.
Software complexity is likely to compete with silicon complexity as Systems-on-a-Chip embed not kilo-lines-of-code, but millions-of-lines-of-code. Software differentiation is increasingly more critical to the customer. Simultaneously, the number of defects in 50 million lines of code is infinitely more than what one would find in 50 thousand lines of code. This requires an entirely different order of software development methodology, test and release discipline.
There will also be increasingly more analog and RF sections on the SoCs in this new world as the order of integration increases rapidly. This will call for significant innovation and a combination of process technology, circuit design and
system know-how to win.
This change from System-on-a-Chip to Systems-on-a-Chip is more than a challenge to engineering scale and complexity. It also fundamentally defines who can and who cannot play in this new arena. Few companies can play in the System-on-a-Chip world—even fewer can play and win in the Systems-on-a-Chip world. The reason is that not many companies will have, or have access to, each of the varied “systems” that are a leadership products in their own right, and which comprise the new Systems-on-a-Chip. Few companies will have the R&D strength and depth to cost-effectively scale this entry barrier, especially at advanced process technologies?
PartnershipsThe skill implications for the future, as these mega trends take shape, are daunting. It calls for engineers to have significant domain knowledge, increased system know-how, and an understanding of the product as a whole—not just the chip design or the software—and the interactions amongst the systems that comprise it. It also implies a need to have a
solid breadth in understanding across silicon and software, across the analog/RF and digital world, to develop leadership and innovative products.
Compound the skill implications with the spread of the customers around the world—each with vastly different needs—and it becomes clear that the challenges cannot be overcome by just one team; this requires leveraging partnerships inside and outside a company.
Partnerships inside a company—leveraging skills and strengths available across multiple groups and locations—can be very powerful. This requires working through and appreciating cultural differences around the globe. Leveraging strengths of partners to supplement, and more importantly, complement one’s strengths can be invaluable too. However, this is not simple either and requires a deep commitment on the part of all players to succeed. To be healthy and sustainable, this necessitates the partners to build win-win relationships, and go the extra mile to support each other.
ConclusionThe three mega trends described here have raised the entry barrier for successful semiconductor companies. Innovation has been redefined in that execution success now has a new meaning. This has raised the bar of excellence for engineers. Today, leaders in this profession must have the ability to scale these barriers and overcome them, yet have the humility to continuously learn and leverage others around the globe. At Texas Instruments, we deeply believe in these values.
Author information You can reach Dr Biswadip Mitra at
media@india.ti.com
