| |
|
|
 |
| |
 | |  | | | | | Interview: Design Data Management Steps up to the Podium
Britta Krueger with Chris Goldstein and Rick Stanton
Qimonda, IC Manage, MatrixOne | |
| |  | |  |
Britta Krueger, CAD engineer, Infineon/Qimonda; Chris Goldstein, Director of Professional Services, IC Manage; and Rick Stanton, Director Industry Solutions Strategy for Global Semiconductor, ENOVIA MatrixOne will moderate a new forum on IC design data management. Launched on July 30, the new community forum serves as a round-table to discuss and share ideas about … well, let’s let them tell us.
cdnusers: Why is there a need for a forum on data management?
Britta: Today, it’s very difficult to manage design data because teams work in different locations, in different time zones. To make it work, the entire design process needs to be organized and communicated—it needs to be verifiable at one point in time so, at the end of the day, designers know which versions of IP block, internal block and external IP to use in their product. You can only do that with effective design data management—and that is very challenging.
Chris: I couldn’t agree more. When a company grows from small groups working in one location to larger teams working in multiple locations, the question becomes, how do you coordinate the work? How does a company enable team members to work at the same time on the same project? What are the logistics? What are the tradeoffs? This is really what we mean by the term design data management.
Rick: I would also add that the challenges of electronic product development require that companies manage the whole product life cycle. Everyone—managers, design leads, designers—needs more visibility into what makes up a design, what trade-offs have to be made, what type of technical problems have to be solved and what tools are actually used. Design intent has to match up with design execution; issues and changes have to match up with logic and layout. It’s all tied together. You need good design data management processes for that to happen.
cdnusers: The problem has been around awhile, why do you think the time is right to create a new forum when, frankly, everyone is short on time?
Britta: Short answer—chips are bigger, product life cycles are smaller, no one can afford to completely design a product from scratch every time. Today, the design task is one of being able to manage IP, to quickly assemble it into a new version, to take existing chips and make derivatives. This is my definition of design data management—empowering designers how to make derivatives efficiently.
cdnusers: This implies creating derivatives today is a problem.
Chris: It is. When engineers are working between different sites, the simplest—but most problematic—way to share date is for site A to make a copy of everything that site B is doing and then start modifying it. Then, when they are done they send the design back to the site A. Unfortunately, team A is stuck trying to figure out what has actually changed in the design. Using this process companies do get chips out—often without errors—but the amount of time it takes to coordinate that effort is expensive in time lost. Newer methods enable incremental updates.
cdnusers: Sure, things are getting more complicated, but aren’t today’s engineering graduates up-to-speed on the latest in design data management?
Britta: No, especially young graduates who are coming out of universities. One of my first interview questions is always, do you know what version control is? They look at me blankly. Engineers with experience, of course, do know about version control. They have learned, because the real discipline of design is in producing product and taking it to market. This is the experience we hope to tap into in the data management forum.
cdnusers: What do you think participants will gain?
Chris: The problem with data management today is that, although there are a lot of different ways to do data management, there are really no commonly agreed upon best practices. In the new forum, we hope designers will ask "can you really do that?" or “how are you doing it?” and share ideas about what’s working for them and, eventually, best practices will emerge.
Rick: I hope through the forum we will open discussions about the benefits of seeing the big picture—that with solid data management practices, design work can be done more efficiently and faster across the entire design chain.
cdnusers: Final question. You are all willing to spend a lot of time as unpaid moderators for the new community. Why?
Britta: That’s easy. Data base management issues have been part of my life as a CAD engineer for the past 12 years. I am passionate about doing things right the first time. I believe that companies can benefit from a good database management solution—that they can become adept at it and make their jobs easier and, personally, they can become more creative and productive.
Chris: Easy for me as well—because of my background, I think I can help with discussions. Before I joined IC Manage, I spent 12 years in the Cadence engineering services department. I learned that design data management, especially for larger chips and multisite design teams, is often a limiting factor in a team’s ability to be successful in having a predictable design process. In the chaos of tapeout—when it’s critical that processes converge—designers do not always have a good way to keep track of design data.
Rick: I’m with Britta and Chris on this. Starting with my time at Viewlogic, and then with Synchronicity both before and after it was acquired by MatrixOne and later Dassault Systemes, I’ve been working with a great team of people who have always believed that the next level of incremental gains in IC design will come from how we collaborate and manage the flow of data between teams. The innovation comes from the needs and vision of people who we hope will be very involved in this forum. I hope to be able to contribute as well as learn from this forum.
cdnusers: Thank you. Where can we find the forum?
The Design Data Management forum is in the Forums Special Topics category of the community.
| 
About the author
Britta Krueger graduated from Technical University in Dresden, Germany with a Masters degree in Electrical Engineering in 1991. She joined the CAD department at ITT Intermetall in Freiburg in 1991 supporting DFII and Design Database Management tools. In 1977, she joined Mosaid Technologies as a CAD engineer with responsibilities in Design Flow and PDK development and support; she become CAD manager in 2000. In 2004, Britta joined Infineon/Qimonda as CAD staff engineer with responsibilities to support Design Database Management and CAD infrastructure topics for a full custom design flow.
Chris Goldstein has over twenty years experience in EDA and related IT infrastructure with several leading companies: he was a software engineer for Calma Corporation, Director of IC and PDB R&D for Electronics Products for Applicon/Schlumberger Technologies, Design Center Director for Zaiq Technologies, and Director of Operations and Professional Services at Magma. He spent ten years in the Cadence Engineering Services group working in mixed-signal IC design methodology and device modeling and complete outsourcing of multichip designs. Chris has a BS in Math from California Institute of Technology and an MBA from San Jose State University.
Rick Stanton is director of Industry Solutions Strategy for Global Semiconductor within the High Tech Vertical Business Unit for ENOVIA MatrixOne. He has over 20 years experience in EDA and Collaborative Design Methodology focusing on IC product design. Prior to that, Rick held a number of Applications Specialist, Development and Support roles at Viewlogic Systems, Racal-Redac and RCA/GE/Harris Semiconductor. Rick has a BS degree in Electrical Engineering from Rutgers University. |
| |
This content has been rated 5 out of 5 by other users |
|
|
|
|
|
|
|
|
|
|
| |
|
|
|