by Marcelle Douglas, National University
Balancing good business with good design practices to protect future growth doesn’t come without its challenges. Electronic designers previously maintained a competitive advantage by recognising when the industry was about to shift, and by being smart businessmen first before being developers and systems architects.
While technology innovation was driving the evolution of the design process, it still had to become widely adopted before it could be mainstream technology. What’s changed is the level of complexity brought on by all of this innovation. We’ve got more sophisticated electronics products, and an ever-shrinking time-to-market window. There’s mounting pressure not only to cope with this increasing complexity, but also to find smarter long-term solutions to keep that competitive advantage and secure growth.
RE-EVALUATING OUR APPROACH
Since the microprocessor, a slew of advances has made things smaller and faster, leading to increased design complexity. As a result, designing electronics isn’t so straightforward anymore—it’s now a question of innovation in the face of managing ever-increasing levels of design complexity.
Unfortunately, electronic design and development tools haven’t experienced the same unprecedented rate of change. Yesterday’s solutions aren’t enough and the benefits of using current design methodologies are reaching the point of diminishing returns.
The danger is that managing the complexity has pushed our focus away from innovation. This effect is compounded because the current device technology in addition to market demands combine to move us beyond the scalability of today’s design methodology and tools.
Exploring available options leaves us with few choices, most of which don’t rely on securing any form of future growth or potential. Raising the feature sets in existing tools only make the integration process more complex and difficult. Hiring engineers with the right skills is expensive and time-consuming.
A re-evaluation of approaching electronic development is overdue. Perhaps the biggest barrier is in our minds and how we perceive the difference between hardware and software.
IT REALLY IS ALL ABOUT
Solving complex development problems lies with raising the abstraction level that board-level engineers work at, which means hiding the levels of complication from the designer. Larger systems can then be tackled in new ways without extending development time in the same manner that ICs, for example, reduced the complexity at the board level and higher-level programming languages simplified development for software engineers. What we’re seeing, then, is a change away from a hardware focus to more of a software focus, and the timing couldn’t be better.
Designs aren’t based solely on hardware anymore, as the boundaries between hardware and software become increasingly integrated. Low-cost, highcapacity FPGAs could change the way we design. They would allow physical hardware to become programmable.
This “soft-design” focus in electronic product development makes a lot of sense. By separating the device intelligence from the physical hardware into which it’s programmed, you can avoid many of the trappings long associated with a hardware- dependency solution. This would include the inability to easily make changes late in the development cycle.
Software combined with hardware then becomes an intrinsic part of the new unified design paradigm. Reconfigurable hardware platforms are driving the momentum for redefining the electronic development paradigm. Thus, they’re accounting for increased interest in a “soft design’s” potential.
TIME TO SEE HARDWARE
IN A NEW LIGHT
Such a methodology would let us reap much more from existing technology. All that’s needed is a suitable design environment. Until now, the perspective of electronic design has been fragmented and focused on the device; thus, development tools evolved to follow suit.
In turn, designing the average printed-circuit board has been based on the choice of processors or FPGAs that reside on them. For example, try improving system performance using the conventional hardware approach and tools without also including the unavoidable and painful process of rewriting lowlevel code or implementing more efficient algorithms. It’s both time-consuming and costly. But until recently, designing the hardware platform (that is, combining all of the pre-built circuits, including microprocessors and logic chips) was a separate process from creating the actual device intelligence.
Continued on Page 2
Adopting a “soft” approach offers some serious advantages, such as more complete design synchronization, design reuse, and a unified methodology. Higher-level abstraction and design automation can be effectively harnessed. That’s because designers are able to mix their approach to hardware and software without fabricating hardware to support the design functionality, and without having to think about it as much.
There’s no longer a need to “fix” the design of a hardware platform before beginning software development, and no longer are countless hours spent trying to make the smallest changes within a rigid development cycle. Instead, you can employ an iterative approach to design, one that allows for experimentation and exploration of “what if” scenarios without increasing design time. Focus, effort, and resources can ultimately shift to higher-level design activities, wherein lies the value. All that’s needed, then, to make this development process a reality is the right design environment.
DELIVERING A VENDORINDEPENDENT
Most of the solutions available today are vendor-specific point tools that tend to be underexploited because they’re not deployed as part of a unified, integrated design flow across an organization. Electronics design could certainly benefit from new thinking, such as incorporating mutual-interest technologies into one single design environment instead of individual ones.
Such a shift would significantly boost mainstream development by allowing all aspects of electronic product development to be designed and managed within a single system, speeding the transition from design to production. New and innovative design possibilities not possible with conventional point tools and more disparate design flows could be considered.
With a more unified approach, designers can focus on higher-level applications, and reuse both their existing work and third-party technology without sacrificing innovation or increasing design times. To make this concept a reality, a design system and the infrastructure to support it must exist. Moreover, there must be a willingness to redefine our view of the design process.