At the end of March Mentor Graphics announced a technology launch of subsystem intellectual property, beginning with the industry’s first USB subsystem solution from a single-source EDA provider. I had a chance to interview Bill Martin, General Manager of the IP Division, and Rick Tomihiro, Director of Marketing for the IP Division, about the change in strategy.
Brief Bio from Rick.
Rick: I started my career as a CPU designer at Amdahl. I continued that at STU Computer Research and then at Tandem Computers. I developed an X86 processor at Chips and Technologies. Then I switched into a marketing function. I was Director of Marketing for NexGen which was acquired by MAD. Then I went to work for Synopsys for 8 years and ran IP marketing. I left Synopsys three years ago to run sales and marketing at a design services and fabless ASIC company. About a year ago I joined Mentor.
What caused you to come back to an EDA vendor?
I have a lot of experience in IP. I like working in the IP market. I knew Bill from Synopsys. He recruited me into Mentor. I think it was a good decision. I am having a lot of fun here.
Bill: We are setting new paths and directions which is always fun to do, to go off and change how the world sees a solution. That can be a lot of fun.
When did you take over the IP division?
Bill: I took it over in October 2005. When Mike Katz left Mentor, I was asked to come and take over the group.
What did you do before that?
My entire career or just at Mentor?
I got a BS in Computer Engineering from Illinois. Got an MBA from the University of Texas at Dallas. I was at Mostek for five years doing microcontroller and memories. After that I went to VLSI Technology. Ten of the years were in the design centers working with lots of customers, doing a very successful business, For the last two years at VLSI Technology I worked for CTO Bob Paine. We did all the strategic planning for the group. Anything from process packaging, libraries, design kits. You name it, we did it for the entire corporation. I left there after 12 years and went to Synopsys. There I rolled out PrimeTime, Chip Architect and Formality and killed the Motis product. I had some friends that had come to Mentor and they recruited me to come here. I was inside the consulting division and ran it for 5 years. Then I was part of the sales team for a year and half working on a lot of internal logistical errors and problems we had to get fixed. Then they said go and take care of the IP group. So I came over in October 2005.
I understand that the embedded software group is being merged with the IP group.
Actually, we just announced that internally. We merged the two teams about 4 or 5 weeks ago basically around this kind of reasoning. We came up with this about 9 months ago. We have been working on it. The USB is the first one that we got fully certified. It just made a lot of sense to merge the two teams. It’s all embeddable IP whether it is software or silicon IP. It makes a lot of sense to get them together.
If you look at the embedded system IP portfolio there is just a real strength the between anything form development tools to platforms to OS to middleware to silicon IP. You can start to connect them and pretty quickly create some pretty large powerful solutions for our customer base. We just think there is a wealth of products that we can get to our customers that can make them extremely successful and get them to market a lot faster.
Presumable you had these separate products before. So is it bringing the test and integration parts together that is new?
Up to around 9 months ago, it may seem strange but the two divisions did not do a lot of work together. We did some work but not a lot. When we started working on these subsystems, it really crossed over the line on both sides right to where we had to work very closely together in the same lab and do a lot of testing to make sure things worked.
This past year we have had pairs of customers, the silicon supplier and the end product maker, probably about 8 really large customers, come to us. They were all having problems between middleware and the controller. They would say that our controller has a bug and we have got to fix it. We said that we were not so sure. We had done a lot of quality checking. We’ve conformed to the standard, passed certification, all of that stuff. We would like to replicate the problem. They could not give us a simple test to show us. Some of our customers brought in small test systems, small boxes that they used in their development and verification. One brought over their box. It took them probably a day to get it back up and running due to the jostling that it took in the flight here. Once they got it up and running, it took literally about an hour for our team of guys to see what was wrong. We told the customers here’s the problem in the middleware. It is not doing the right interfacing with the controller. The customer said that that is replicated throughout the code. But we can go off and change a couple of instances and see if it gets past this point and works. They came back that day and actually retested. It got past the problem point. They said they would go back home and fix the rest of the code. We never heard from them again. We know that they launched their product, a high demand consumer product, into the marketplace and they were successful. We have seen that experience replicated by a lot of customers. We said we have a unique opportunity here. We happen to have both components inside Mentor. Why not put them together, fully test them and make sure they work, are certified and supported by one vendor, namely us, rather than have customers buy from multiple vendors and point fingers at different vendors saying your stuff is not working? We can cut all of that stuff out. We can make sure that everything works and is fully certified. Customers can buy it from us, get support from one place, and accelerate their development time. They don’t have to fix these problems or develop solutions in this area. They just buy it from us.
How big were the two groups before the merger?
We are talking about a combined group of 250 to 300 people. I do not know the exact count. We have development groups around the world. There are probably eight different sites between the divisions. It goes from Cairo to the US to Pakistan.
Is the combined unit 50% semiconductor IP and 50% embedded software?
It is very similar to what our customers are seeing. There is a lot more on the software side. Don’t forget on the embedded side we do not just have IP like the Nucleus cell lab, we’ve got the Inflexion Platform, the developer tools area. We have people working on other tools not just embedded IP. This includes platforms to help you develop your system faster but it is not true IP that you put inside the chip. I do not have the breakdown on the number of people in the EDGE area or on the Inflexion Platform or the Nucleus.
On the revenue side how much did the groups generate?
I don’t have the final numbers. You would have to get that data form our industry intelligence person. (He would not give me the data.)
Does a company like ARM compete in this area?
It is interesting. They have a processor and might have some of the PHY but I have not seen much in the middleware area or any of the controllers.
Rick: They do not have middleware. They have a couple of PHYs for PCI-Express and such but they do not get into the controller space. It does not make sense for them to do any middleware in the IP space.