Convergence: The new design dilemma

Consumer demand for convenience is giving an impetus to the implementation of convergence solutions. Cadence CEO Ray Bingham looks at this mix of communications, computing and consumer electronics and the design challenges facing companies trying to keep up with the market economy’s insatiable hunger for increased computing power, speed and the need for multi-functional devices

I bought my first wireless telephone back in the 1980s–a big, bulky device lugged around on a shoulder strap. Costing many hundreds of dollars, it let me do just two things: Place a call and receive a call. It was, in the truest sense, a communications device.

Of course, we know what’s happened to the wireless telephone in the ensuing decade. New generation phones cost significantly less and fit neatly into a shirt pocket or clip inconspicuously to my belt. The most important change, however, is what they can do. Yes, we can still place and receive calls, but that’s where the comparison ends. Today’s wireless phone can be used to keep appointments, browse the Web, ‘chat’ with associates and friends, and even play games.

In other words, it’s no longer just a communications device; it’s a convergence device—one that mixes features and capabilities from the once-distinct worlds of computing, communications and consumer electronics to provide greater value to the user.

We are surrounded by examples of how convergence devices are changing our world and our lives. In the medical world, paramedics in Africa can diagnose a rural patient using a handheld ultrasound device and share the information with medical experts in an urban hospital a thousand miles away. In the automotive world, a service agent can bounce a signal off a satellite to pop open my car door should I lock my keys inside. Even in the once-mundane world of household appliances, manufacturers are beginning to embed digital intelligence and communications capabilities into refrigerators, dryers and dishwashers that let technicians pinpoint failures before leaving the service centre.

None of this convergence would be possible without recent dramatic increases in silicon capacity that have led to innovative developments at opposite ends of the customer spectrum. In the consumer world, the economics of silicon capacity enable companies to design processor-based products that meet consumer markets’ stringent price, size, power, and reliability demands. In the infrastructure world, whole rooms of networking equipment are being replaced by modestly-sized boxes, such as gigabit routers that have given way to terabit routers.

The new design dilemma

Behind these advances, however, is a dilemma. Despite the escalating complexity posed by shrinking geometries and convergent worlds of computing, communications and consumer devices, the process of electronic design remains largely unchanged. It is still partitioned into four distinct and separate design ecosystems—emb-edded software, digital logic, analogue circuitry, and printed-circuit boards–that are assigned to teams of engineers with specialised and disaggregated knowledge, languages and tools. Little if any interaction occurs between these ecosystems until late in the process, often so late that expensive and time-consuming rework is required to coax the desired performance from the electronic device.

Moreover, the complexity is increasing. In the consumer world, for example, there is escalating demand for embedded wireless communications, which calls for increasing interaction between the digital and analogue worlds, not to mention an insatiable appetite for the embedded software required to create differentiating features and value. And in high-performance applications, new devices have so much functionality that they need thousands of pins to interconnect to the rest of the system, often creating bottlenecks affecting system performance

Meeting this challenge head-on requires a fresh look at how we approach the discipline of electronic design. It requires a vision of design chain management that—similar to how the supply chain management addresses the movement of tangible goods from concept to consumer—defines a process for the sharing and movement of technology, intellectual property, knowledge, and design capabilities across each contributor to a design.

A design chain would, in effect, concentrate on the management of the white spaces between the unique design ecosystems, focusing in particular on three areas of design convergence.

Hardware-software convergence

Unlike traditional IC designs, new generations of devices are rapidly becoming more of a software effort than hardware. In fact, the amount of software engineering content in such devices already exceeds the hardware content, and the percentage is growing.

As this rich mixture of hardware and software grows, so too must the interaction between the hardware and software design teams. They no longer can rely upon a loose handoff based on individual experience, judgement and rough calculations to partition the design into independent worlds.

Breaking down the barriers between the hardware and software design ecosystems has enormous implications for IC providers and their systems’ customers including the adoption of new IC platform strategies, tools and processes that enable rapid co-development while protecting massive software investments.

Digital-analogue convergence

The insatiable appetite for communications capabilities in digital devices is creating explosive growth in mixed signal ICs. Within the next five years, analysts predict that nearly three-quarters of IC designs will be mixed signal.

In much the same manner as the hardware and software design worlds, however, companies have isolated analogue circuitry from digital circuitry for a number of reasons. Analogue circuitry is highly sensitive to noise, while digital circuitry is relatively resilient to it. It is much easier to design digital logic if the digital designer does not have to worry about analogue requirements such as extra supply wires and shielding for sensitive analogue circuits—especially since they are generally unfamiliar with analogue circuitry on the chip. And certainly it is easier for analogue engineers to design analogue circuitry without having to worry about the effects from digital logic.

The design approaches are completely different as well. Analogue design involves the frequency and continuous time domains, whereas digital design is event-oriented and centred on clock cycles. As a result, the designer skills are completely different with analogue designers requiring more mathematical skills and digital designers requiring more computer architecture skills.

As mixed-signal development increases, new top-down approaches must be developed that solve the issues related to packaging, flexibility, noise, yields and testing.

Silicon-package-board convergence

Exponentially increasing silicon capacity and performance affects more than just the silicon; it also affects the IC packaging and printed circuit boards (PCBs) that interconnect the silicon in a complete system. There is little value in IC advances that keep pace with ‘Moore’s Law’ if the packages and PCBs that use them cannot.

New levels of integration are required as pin counts exceed 1,000, digital frequencies exceed 500 MHz, and sensitive analogue signals cross IC boundaries. These factors are creating explosive growth in custom packaging, which—over the next five years—will grow to nearly 16 billion, up from four billion today. In other words, the days of standard packaging are numbered.

The design chain

Creating a design chain represents an enormous challenge to the electronic design industry, but it also creates an attractive opportunity for companies that aggressively tackle that challenge. By some estimates, more than 50 different tools from several different companies are used in designing today’s electronic devices. As devices become ever more complex and the electronic design industry continues to disaggregate and specialise, that number will grow. Design chain management will require co-ordination and co-operation on a number of fronts to meet the needs of systems customers.

First, new generations of technology and tools that work together must be developed, enabling a holistic approach and ignoring the artificial boundaries of distinct ecosystems. Further, these new technologies and tools must achieve a higher degree of openness so that information can be shared across the design chain.

Second, we must develop better ways of sharing information across the design chain. Given the technical challenges that our customers face, closed architectures no longer make sense. They cannot afford the lost time or investment required to make the output of one tool work with another.

Third, new services must be developed that focus on the delivery of customer solutions. Great tools are not enough; as today’s customers struggle to tackle the challenges of convergence, they are hungry for our knowledge. This represents an enormous opportunity for the electronic-design world to offer value beyond the intrinsic tools by delivering total solutions.

Fourth, closer partnerships across the design chain must develop so that higher levels of interoperability can be achieved that reduce time and lower costs. Given the complexity of today’s design challenges, no single company can satisfy all requirements.

But these are baby steps compared to revolution to come. Even the current economic equivalent of a hard reset cannot and will not stop the advance of technology. History teaches that while economies may rise and fall with the times the pace of complexity and the yearning to innovate is constant and growing. The electronic-design industry stands to play a crucial role in triggering the next wave of growth, but only if we understand and solve the challenges of design convergence. By closing the gaps on traditional methods we will be making it possible for the electronic-design world to get designs completed in time to meet tight market windows and, more importantly, at a reasonable cost.

The simple truth is that market convergence is an unstoppable force. The hard truth, though, is that electronic design in this convergent world is tough to do because of the simultaneous convergence of the once-distinct ecosystems. Only by broad participation of companies across the design can we help our customers cope with the challenge of inventing the next generation of innovative electronic products.

Ray Bingham is President and CEO of Cadence Design Systems Inc., provider of electronic-design products, methodology services, and design services to the wireless industry.