August 2, 2005 -- For this month's SOCcentral Executive Interview, Contributing Editor Jim Lipman sat down with Mike Kaskowitz, vice president of the IP division of Mentor Graphics and the president of the VSI Alliance, the open organization developing SoC, IP and reuse standards to enhance SoC design productivity. Mike talked about the effectiveness of the Alliance's Pillar structure, instituted in 2004, relationships with other standards organizations, and the future role of VSIA.
Lipman: Last year, the VSIA changed its structure from one of several working groups, each with their own agenda, to a Pillar structure, with each of the four pillars focusing on a particular IP issue. How has this worked out?
Kaskowitz: I think this has worked out very well for us. Basically, the Pillar structure supports an organization whereby the VSIA board of directors sets the strategic direction for the organization but is not responsible for driving the standards activities. What we do is create a team called a Pillar, which has marketing, engineering and business functions. They are responsible for reaching agreement on what a standard will be for us. By going to this structure, we actually are able to move very rapidly, which has always been an issue for a larger standards organization such as the VSIA. Just recently we released our second revision of our Quality IP (QIP) metric and we have put out a preliminary version of our IP Protection standard as well.
Lipman: You mention that you have faster response and faster development under a Pillar structure. What are the other benefits you're seeing so far?
Kaskowitz: I think it helps foster "buy-in." One of the issues, historically, is that we would technically create a standard and then just throw it out in the industry. People would use portions of it, but would not really embrace it as a standard. The Pillar allows us to focus and gets like-minded individuals on a particular topic together from marketing, engineering and business standpoints to really work on issues. A lot of times, the technical aspects of an issue do not address the practicality of its acceptance and implementation. People can work on gaining consensus of the Pillar members. For example, prior to our release of the QIP metric 2.0 we spent about four months in alpha review with about eight or nine of our larger company members and they all dedicated resources to deciding what would be feasible from a business standpoint as opposed to technically what's feasible.
Lipman: You have had a working relationship with the Spirit Consortium, the two-year-old group creating standards for describing and handling IP. What is the status of this relationship and do you currently have strong ties with any other standards-type groups working on semiconductor IP or the tools needed to handle this IP?
Kaskowitz: We have a very good relationship with Spirit. Last year, Ralph von Vignau, chairman of the Spirit Consortium, joined the VSIA board of directors. The reason that Ralph joined was that we both envision that there are a lot of areas where the VSIA and Spirit can actually work together. Spirit has been focused a lot on IP interchange standards and the packaging constraints whereby the VSIA has been focusing on issues such as IP quality and IP protection. Just recently, VSIA formed an IP Transfer Pillar, which is working very closely with Spirit since both organizations are working in the same space.
As far as other organizations, the biggest standards body we work with in this area is the FSA [Fabless Semiconductor Association]. The FSA has always done a wonderful job with arranging the marketing activities and speaker events for the fabless industry; thus, they speak in a political fashion for the industry. VSIA, on the other hand, has always been associated more with technical standards. In recent years, we have been working closely with the FSA where we would drive technical activities and take advantage of their marketing prowess to promote our activities to the world.
Lipman: Any specific areas in which you and the FSA are working together?
Kaskowitz: Yes, in fact on the QIP metric. The initial quality IP metric was a soft-IP metric. We have been working recently with the FSA on a hard-IP version of the QIP metric, since many of the FSA members are very familiar with hard IP.
Lipman: The current VSIA pillar structure provides a means for the larger and more established electronics companies to better support ongoing VSIA work. This also allows them to more easily "push" their ideas through the Alliance. How do the smaller VSIA members get their thoughts across?
Kaskowitz: What we have is a membership structure that allows any company to become a principle VSIA member. The old VSIA was set up around the concept of company size. The larger the company, the more dues they paid and the more input they had. This is no longer true. The organization now still has a dues structure based on company size but, in addition, the voting and the power and what gets worked on in the VSIA is determined by principle members, regardless of company size. If somebody is willing to contribute and sign up resources to the various Pillar activities, if you want to put people on the board, the VSIA won't constrain you because of size. Practical issues will, however. Since it is more difficult for a very small company to have a large presence on all the Pillars they have to pick what Pillar activities are most important to their business. On the other hand, the larger companies might have representatives on all Pillars.
Lipman: Of the four current Pillars, three deal with specific IP issues-Quality, Protection, and Transfer. The fourth, the R&D Pillar, appears to be a "catch-all" to investigate other IP issues that may, eventually, be deemed important by the VSIA. What determines the topics on which the R&D Pillar works and what is the mechanism for moving evolving projects to, possibly, new Pillars?
Kaskowitz: You're right, Jim. The R&D Pillar was started as a "catch-all." We were initially focusing our short-term activities in quality, protection and transfer. We have talked about in recent months and current thinking is that best usage of the R&D Pillar is to do extended research in the same areas as we have in other Pillars. So, rather than be completely [concerned with] random research the thought is to actually focus in the areas of quality, protection and transfer, but on topics that are not on the immediate horizon, like three years out or so, but also may never reach fruition. We are trying to see if there is a way to better align the R&D activities to the activities of the sanctioned Pillars.
Lipman: Can you give me some examples of what might be extensions of work being addressed in the other Pillars?
Kaskowitz: For example, we have an IP Protection Pillar and they are working on the easier objective of IP tagging, how you put traceability and tracking into the process. A more grandiose solution would be to come up with a physical encryption scheme and get the tools providers to modify their tools to support this encryption. That's a longer term prospect that may not happen, so an interesting topic for the R&D Pillar to drive is can we get some agreement on what an encryption standard for IP protection might look like.
There are other follow-on areas for IP transfer. Right now, we are working on some of the simpler mechanisms of transfer. As it evolves, you have to start worrying about how you can transfer the verification test benches, how do you transfer the documentation, how do you value the IP that's being transferred? There are areas of activity that encompass larger problems to solve and they are not as immediate.
Lipman: This doesn't mean that you are going to give up on some of the currently addressed issues by the R&D Pillar such as analog firm?
Kaskowitz: No, not at all. Once again, the difficulty is working in areas where there is membership interest and that members will drive. What happens in trade organizations a lot, which was surprising to me when I became the VSIA president, is that people love to suggest ideas; they don't necessarily want to put bodies on getting them implemented. The nice thing with our Pillar structure is that a Pillar does not exist unless there are enough bodies committed to it by member companies who are interested in it. To some extent, while there might be an activity going on in the R&D Pillar it's like a Birds of a Feather session at a conference. If you get enough Birds of a Feather that have an interest in an area, then we start an activity in it. Otherwise, you probably just have a couple of engineers who like to talk about a topic. And that's all right, too - they have the mechanisms through our website, email and meetings to have that conversation but it's not likely that a real standard will come out of it if it doesn't have some large resources from several companies.
Lipman: What do you see for the VSIA five years from now? In other words, will it still be a separate organization for developing SoC, IP, and reuse standards or will it have disappeared or become part of a larger group addressing IP issues?
Kaskowitz: I think that the VSIA has positioned itself and is perceived as the umbrella organization for IP and design reuse. I also believe that we have too many standards bodies. I do think there is going to be consolidation with the standards bodies, so I think that VSIA is in a good position to be that umbrella organization. I would envision, over time, that some of the smaller standards bodies or standards activities would probably roll into the VSIA.
Lipman: Do you think the VSIA structure will stay pretty much the same as it is now?
Kaskowitz: I would think so. We have had the structure in place for about 2-2 ½ years. The Pillar structure is kind of an extension of what we did earlier when we massively changed how the VSIA was structured. I like to associate the VSIA structure in terms of the U.S. government. There is a centralized government, which is the board, and decentralized governments, which are the Pillars. What we did is push down the entire decision making and the empowerment into the Pillars-that was the big change. It worked out wonderfully and also sets it up for consolidation with other organizations. They would come into the VSIA under the Pillar structure and would put a board member on the board to help set future direction, but the operations of their team that is already working in an area on a daily basis would not change.
Michael Kaskowitz is the general manager for the Mentor Graphics Inventra IP Division, and serves as President of the VSI Alliance, the leading IP standards body. At Mentor Graphics, he is responsible for driving the company's strategy to deliver complete system-on-chip (SoC) solutions through the application of hardware and software IP. Kaskowitz brings 24 years of technology and management experience in the IP, embedded software and EDA industries to Mentor Graphics.
Prior to joining Mentor Graphics, Kaskowitz served as vice president of engineering at Sensory, Inc. and, earlier, he served as vice president of engineering for Cadence Design Systems' Central Architecture and Technology, Flow Engineering, Product Engineering and Usability Engineering departments. He has also held positions with Compression Labs, VLSI Technology, Motorola, and Apple Computer. Kaskowitz holds a Bachelor of Science in Electrical Engineering from the University of Illinois, Champaign-Urbana, 1980.
Go to the VSI Alliance (VSIA) website to learn more.