Published on June 15th, 2009
The top two reasons driving organizations to use design data management systems today are tracking and fixing bugs and efficient team collaboration. This input came from 416 IC design professionals that recently completed IC Manage's blind, anonymous online survey.
We heard from digital and full custom designers, CAD managers, verification engineers, software and firmware developers, engineering management and FPGA designers all around the world and have made the results available in a Free Online Report.
Figure 1. Main reasons to use a design management system.
Verification issues are the top driving force for design management – a fundamental verification requirement is to track/close bugs with the exact version of the design data or the specific line of code in the design file. This critical information about which bug and which design file it was fixed in can make or break a successful tape-out.
The second major reason given for using design data management is to enhance team collaboration, driven by increasing design complexity and the work distribution required for today's 24/7 global design. Other significant reasons for deploying design management were easier access to working configurations, improved designer efficiency, better product quality, reduced overall project time, and better IP reuse.
The IC design respondents also pointed out that design management issues create substantial overhead. Design teams spent an average of 12% of their work hours on issues such as tracking down IC-related design files or a specific configuration of files. For a 50-person engineering team with a cost of $10M/year, this 12% additional overhead to designer's time equates to $1.2M annually.
Over half of IC design organizations expect to have a design data management system in place by the end of 2009. Yet the overall respondents still cited two major obstacles to deployment. First, they said the design management process was difficult to implement, and second, they felt the tools were not well integrated.
I would argue that deployment difficulty has stemmed in part from the fact that traditional commercial hardware design management systems carry forward practices that rely on file trees as a way of organizing data, so designers are burdened with worrying about where things are stored or are forced to traverse cumbersome files trees to find what they need, causing them to search endlessly for files and proper versions.
The design management system should be an intuitive part of the designer's existing methodology. The design data is best structured according to the way the design is organized and packaged in "views" that make sense for a designer's particular job function and workflow. In most cases, an individual will be assigned to a particular stage of the design, such as schematic, RTL or layout. The design management system should be flexible enough to only present them with the data they need – this will simplify their task as well as prevent them from accidentally modifying data outside their scope. This approach addresses both designer overhead and deployment problems.
The historical lack of integration of design management has also limited broad adoption of design management, making design management requirements a cumbersome 'overhead' to designers. Instead, IC design management, including revision control and configuration management, should become a transparent part of their design flow. The design management interface should model the look and feel of the designer's default interface, preserving commonly used features, yet allowing additional design management functionality such as browsing through libraries, check-in, check-out, viewing version history, and design synchronization. The system should let designers automatically track and fix bugs between different design versions and different design tools.
We've come a long way in the 15 years since the first commercial design management tools brought the communication and collaboration capabilities of the Internet to IC design teams. Outsourcing is now common, and we have round 'around-the-clock' design work done globally. Design management's primary value is to improve collaboration, not just within the design team, but across this disaggregated global design chain.
Entire enterprises are seeing increased efficiency by transitioning from a collection of disjointed version control and configuration management approaches to a unified design management solution from concept to manufacturing. Emailing files, sending terabytes of data across FTP servers, or sending data files via overnight mail are no longer sufficient.
42% of CAD Managers stated their organizations had actually missed a project deadline or delayed a tapeout due to design management issues such as version control or configuration management. The average length of the delay was estimated to be 3 work weeks, increasing costs by $600K for a 50 person engineering team.
To survive in today's economy, organizations must look under every rock to come up with ways to improve their design and verification process. There is no room for error or delays in this unforgiving market.
Dennis Harmon is Vice President of Business Development at IC Manage. He has more than 20 years of management experience in the electronic design automation industry. Most recently, Dennis was CEO of Zenasis Technologies (now Open-Silicon). Prior to Zenasis, Dennis was co-founder and CEO of Synchronicity (now Dassault), a design management solutions provider. Prior to Synchronicity, he held various corporate management positions at Viewlogic Systems (now Synopsys), including heading up the company's worldwide strategic accounts and corporate applications organizations. Dennis received his BSEE and BSME in Electrical and Mechanical Engineering from the University of Maine.