The Internet and EDA: survival of the fittest

Electronic News, Jan 13, 1997 by Mark A. Miller

Boston--For most of us, it's been there all along. What's the big deal? As charter members of the electronic design process, the EDA community has been living with networked computing and electronic communication for years. After all, the Internet was a by-product of DARPA (Defense Advanced Research Projects Administration), a catalyzing agent for many facets of the semiconductor and tools industry.

But, something has changed. It's no longer about E-mail and network access to data and compute resources. While the EDA industry has been focusing on solving difficult algorithmic problems for synthesis, verification and physical design, three forces conspired to "change the game."

* Design complexity has exploded, creating an ocean of complex, interrelated design data and methods.

* Time to market/profit windows for new design projects have collapsed as a fiercely competitive global market for products dominated by high-volume, high-complexity and high-performance silicon has emerged.

* Strong corporate and consumer interest in networked communication and information access have spawned major infrastructure investments in "connectivity" and high-speed Internet access.

Design teams are now forced to develop and implement new design techniques to address these pressures. Team-based collaborative design and Intellectual Property (IP) re-use are no longer interesting ideas, but compelling requirements for successful design.

Every five years, the ever-expanding needs of semiconductor manufacturing technology drive the EDA industry through tool and methodology paradigm shifts. In the early '80s, stand-alone proprietary engineering workstations offered productivity and communications advantages over manual methods. The advent of standard workstation platforms and desktop design validation in the mid-80s provided yet another paradigm shift and further enhanced design size and complexity. In the early '90s, language-based design and synthesis revolutionized the design process, allowing larger designs to be successfully attempted.

Past paradigm shifts were platform- or tool-related. We are now at another EDA tool and infrastructure crossroads--the era of electronic design data interdependence, where sheer quantity and relational dependencies of design data have become the dominating factor in successful design in the mid- and late '90s.

A microprocessor manufacturer found its new commercially available 200MHz microprocessor design took greater than 4 Gigabytes of interdependent design data to represent fully and also document the design and ensure its reproducibility. Its expectations are that this will increase 6X in the next two to three years. With hundreds, sometimes thousands of capture, validation and implementation interations that must be explored to ensure cutting-edge design success, this data management challenge appears daunting.

The need to support multi-site activity on a single design project further complicates the challenge. Access to relevant design data, processes and methods, and data interdependencies anywhere, anytime is the mandate of the '90s. The Internet and World Wide Web offer a foundation to build the next generation of the EDA industry to address these challenges.

How do we capitalize on these new technologies? Many infrastructure elements needed for Web-based collaborative design are free to the end-user. Corporate network and Web access will be provided courtesy of most companies' information technology (IT) departments. If the Internet, outside the corporate firewall, is to play a large role in IP distribution and support, then data security looms large.

Suppliers must build tool versions that will accept encrypted data as input and then decrypt it internally at run-time to allow IP developers the security they need to make the IP re-use market grow rapidly.

EDA tools must evolve to access design data directly, wherever it resides, without moving it to a local machine for tool execution. Release management and revision control also become critical capabilities in this networked data access environment.

It's impossible to predict exactly what the Internet and Web mean to EDA. Marketing and product information is already there. Communications-dominated applications, like customer support and bug fixes/upgrades, come next, perhaps followed by IP distribution and maintenance.

Implications of the Web and Internet are huge. They represent an opportunity for a new tool and design infrastructure, providing better communication, easier and faster access to design data, improved visibility of relevant project information and design team management.

The Internet and the Web are a new fact of life for the EDA and semiconductor industry. In the true spirit of Darwin's natural selection, the EDA evolutionary clock is ticking. Only the adaptive and fittest will survive.