Lip-Bu Tan, President and CEO, Cadence Design Systems Inc.

Earlier this year, to stay on top of end markets, I traveled to Las Vegas and spent two days walking around the Consumer Electronics Show (CES). As I did, I thought about my FitBit and how much data it would capture over all the ground I covered. I reflected on the engineering that’s enabled that type of wearable technology in just a few short years and transformed how we think daily about our health.

At CES, I saw so many wearable technologies, both in booths and on people (a lot of Google Glass!). We live in an age in which a small device on our wrist can tell us to sleep or hydrate more. We can watch—in the palm of our hand—live-streaming video of a man leaping from the edge of space to Earth and track how his body is reacting on its way down.

Sixty years of tireless semiconductor, software and systems engineering effort has gotten us here, and will propel us into the future.

Soaring Expectations

But despite these amazing advances, we face continuing and new challenges, both cultural and technological.

Culturally, we have built enormous expectation among society— among consumers—that we will continue to wow them with new technologies, quickly and regularly. We expect 50 billion semiconductors in Internet of Things (IoT) devices by 2020, many of those devices wearable. Cloud computing handles 3.3 zettabytes of traffic a year, and we expect that to continue growing 25 percent annually. In end markets, we expect sales of semiconductors into the cloud and the IoT to grow 54 and 35 percent respectively.

As an industry, we’re confronting challenges we’ve never faced before. Earlier eras were often dominated by large volume, high margin systems we designed toward: Mainframes in one decade, PCs in another, cell phones and wireless infrastructure in still another. Today’s emerging markets are fundamentally different: There are few “killer apps” or systems that are emblematic of the times; in fact, the possibilities for device and systems applications are almost limitless. But for these markets, products for those applications are price sensitive, rely much more on software for their functionality and have short life cycles.

The rise of wearable technology in particular stresses several crucial design considerations in any electronics system:

  • Power
  • Performance
  • Area
  • Reliability
  • Time to market.

Design engineers don’t have the luxury of being able to focus on just one or two of these attributes; they’re required to optimize all of them.

This situation—unprecedented in the history of electronics—has forced a fundamental transformation in the electronics ecosystem. It’s a transformation during which old notions of design planning, implementation and verification are being reconsidered, reformed and revitalized.

This transformation is intended to keep engineering teams fast, flexible and productive. And it’s implicit in the ITRS roadmap: Between now and 2020, engineering productivity (in terms of gates designed) is forecast to double.

Increasing Complexity

How do we get there from here? As an industry—as fabless semiconductor vendors, EDA and IP companies—we get there by enabling the system-design aspirations of companies. These companies increasingly want to differentiate at the highest levels of the design flow: the system level, the software level. It’s where they add value. Application-driven system design is clearly where the industry is headed, and verticalization—which gave way to specialization starting in the 1970s—is making a comeback.

The evidence can be seen in a new breed of systems companies that are hiring teams of semiconductor engineers, when just a few years ago they were buying what amounted to off-the-shelf components to serve their system needs. In some cases, they’re designing and building the complete systems they once bought from vendors. They are optimizing all the way from the system-onchip (SoC) level to the board, package, and system because their business, storage, networking and power-management needs are so unique that that’s the most cost-effective path for them. And they need our help.

They need help because system-design complexity increases ceaselessly.

Consider the design evolution of the Amazon Kindle, first released Nov. 19, 2007. That first version contained a single-core ARM 11 microprocessor, 250MB of storage, grayscale screen, and interfaces for the keyboard and CDMA communication. The Kindle Fire, released just five years later, contained a dual-core ARM A9, 1920×1200 HD display, up to 64GB of storage and interfaces for touchscreen functionality, various wireless protocols, sensors, audio and video cameras.

And in the five years between those two milestones, Amazon released other new Kindle versions at the rate of at least one a year.

That type of innovation and pace requires the electronics-design ecosystem to evolve as enablers and facilitators of engineers’ system design aspirations, to conceive of and deliver solutions that stretch from IP and VIP all the way through to design tools.

To stay on the productivity track, the ecosystem has expanded the use of IP cores to speed design and increase flexibility in the face of changing protocols and market demands. Where a 90nm SoC contained an average of 18 IP cores, today’s 32/28nm SoC uses 91. At 14nm, it will jump to 123 cores. This type of complexity raises design and verification challenges.

To keep us in line with Moore’s Law and “More Than Moore,” the ecosystem is packing more into less, using 2.5 and 3D packaging technologies that mean more design complexity, greater thermal constraints, and pressures to maintain signal integrity. FinFET devices shine a spotlight on timing and signal integrity issues and require a reconsideration of design flows, methodologies and solutions.

This increased complexity is occurring against the relentless time-to-market pressures of markets such as wearables. It’s tried and true: the competitor who gets to market first takes much of the potential revenue and most of the potential profit.

Enabling System Design

In the face of these challenges and to better compete in this environment, semiconductor and systems companies would be better positioned for success if they had advantages such as:

  • Better predictability into whether their design will actually meet performance, power and other design targets. This can help mitigate risks. For example, knowing that a design could be late for the holiday shopping season might prompt resource adjustments, when the issue is discovered, to avoid the delay.
  • The ability to get an earlier start on certain tasks in the design cycle, such as verification, signoff and software development. This can accelerate the entire design cycle, lower overall costs and prevent costly respins.
  • The ability to tune hardware and software components together to better fit the end product. This can also support a faster design cycle. In addition, since embedded software has taken on a larger role in differentiating electronic systems, access to sophisticated software content can help streamline the design cycle. It’s important to remember that today, 80 percent of SoC development costs come from software, verification and validation.
  • A design flow that enhances engineering productivity, even in light of the fact that engineering teams are commonly distributed across global sites.
  • The ability to differentiate without taxing internal resources or impacting the project schedule. Design teams are expected to meet PPA (performance, power and area) targets. Being able to create a product that stands apart from the others on the market in an enticing way—and that reaches the market first—is the Holy Grail.

The Path Ahead

As I look at my FitBit, I imagine how its engineering team will make that little wristband more powerful and even more useful in coming generations; will it make reservations for me at a nearby health-food restaurant after a good workout?

Those engineers won’t be doing it in a vacuum. Successful engineering teams in the wearables market—indeed successful fabless companies in general—will take a holistic view of system design enablement, one that recognizes their need for a broad portfolio of tools, intellectual property (IP), software content, and services that will help them win.

Systems and SoCs have become so complex that it’s pushed us to transform our business; EDA is no longer only just tools or just about productivity. It’s also about possibility; that is, opening new avenues of opportunity for semiconductor and systems designers to reach toward broader, bigger market opportunities and in the process help transform society.