Google FeedBurner is phasing out its RSS-to-email subscription service. While we are currently working on the implementation of a new system, you may experience an interruption in your email subscription service.
Please stay tuned for further communications.
Get email delivery of the Cadence blog featured here
It’s been five years since Cadence acquired Denali Software, and that will be the cause for some extra celebration at the Denali Party at this year’s Design Automation Conference (DAC 2015). To help set some context for this milestone, Brian Fuller and I had a recent conversation with Sanjay Srivastava, CEO of Denali until the Cadence acquisition. In this interview, Srivastava talks about the history of Denali, the Cadence Denali purchase, why EDA companies are providing IP, the future of the IP business, and his new educational startup, Vocareum.
Q: Sanjay, can you remind us when and why Denali was formed and what technology you offered?
A: We started in 1996 with four co-founders. At that time, hardware/software co-verification was a big thing in the industry, so we put a business plan together. As we did that we realized there were too many components. You needed a processor model, a bus model, and a memory model. So we thought we should do one thing at a time and we started doing memory models.
We were thinking that in a year or two, we would start doing something for real, but, as it turned out, we stayed with memory models until the Cadence acquisition 14 years later. Memory models used to be part of a broader modeling solution. But in the early days of ASICs, when customers needed to verify designs, the memory was often the most significant component that was not already integrated onto the chip. At that point, a standalone memory model market emerged.
So, for a few years, we did memory models, and then we did memory controllers and verification IP [VIP]. The major companies were not doing memory controller IP. Cadence started the VIP market with eVCs [e Verification Components] and Denali was the largest player in that market when the acquisition occurred.
Q: From your perspective, what was the original idea behind the Cadence acquisition and what were the results as you look back five years later today?
A: I believe the highest level objective was to establish an IP business. I think that strategy has succeeded. With Martin [Lund] there, Cadence has gone in a fairly short period of time from not being represented in the [IP] market to being one of its largest players. Also, VIP has started to grow and has become an important product category in the EDA industry.
Of course there has been lots of activity since the Denali purchase, including the big acquisition of Tensilica and lots of investments in analog. Hopefully the Denali acquisition provided some of the anchor position, both from a customer and a technology perspective.
Q: Should EDA companies be in the IP business? Why?
A: The charter of EDA is to enable semiconductor design, and IP is one of the big enablement pieces. It got started as part of the move to fabless semiconductor companies. When you’re talking about a digital SoC, there is certainly no way you can design all the IP yourself, so you need to go out and get into the market. There is a massive amount of mission synergy with the EDA companies in terms of helping customers design semiconductors, and a massive amount of channel synergy as well.
What has become obvious is that when you’re moving to 16nm and 10nm, the cost of IP is very substantial. Ten years ago you could build an IP company and have a meaningful impact. Today the responsibility and burden of enablement for a digital SoC has to be borne by large companies. Hopefully the Internet of Things and MEMS will open up new opportunities for smaller companies, but servicing the SoC market is up to Cadence, Synopsys, Imagination, and ARM for now.
Q: Where do you see the IP/EDA industry going forward in the next few years, and what are the challenges? The EDA industry has consolidated and revenue is sort of plodding along.
A: If we go back to the core mission of semiconductor enablement, I think the possibilities are endless. I believe the Internet of Things is presenting a very large opportunity. What are the common enablement components that are needed to go provide that? Instead of having to design the next big SoC in 12-18 months, you might build the next great SoC as some sort of IoT platform.
Q: How long did you stay with Cadence after the acquisition, and what were you doing?
A: I was there for about 18 months. I was senior vice president of the IP and Services Group. Lip-Bu Tan asked me to run the Services Group, and it was actually a lot of fun to see the massive capability we had in analog/mixed signal.
Q: What did you do after Cadence?
A: I tried angel investing. I did six or seven investments. My alma mater, University of Illinois, launched a program called Faculty Entrepreneurial Fellows program and I’m on the board of trustees. If faculty members have good entrepreneurial ideas, they can stay for a year and get the support they need to try it out.
Q: And what are you doing now?
A: I’m spending most of my time in the education market. I’m CEO of a startup, Vocareum. This company is building what is essentially a cloud platform for computer programming classes. The first generation of what moved to the cloud is content delivery. What Vocareum is trying to do is the next thing, which is to move the students’ work to the cloud. So if you’re doing an assignment or a lab you can use the Vocareum cloud platform and all you need is a browser.
Q: What’s your assessment of the next-generation engineers who are about to come into the workplace? Do they have enough flexibility and multi-dimensional skills?
A: If you’re in school right now, you need not only a computing background—you need to understand data. That means you should go out there and understand statistics and probability. The demand on the data science side is so high that in the last three years, computer science enrollment has doubled in U.S. universities.
I’m optimistic. In my opinion, the universities I’m in touch with are doing an awesome job of training the next generation of engineers, and they also have substantial entrepreneurship programs.
- Cadence Denali Acquisition Results in Broad Verification IP (VIP) Offering
- Q&A: Mark Gogolewski on Denali History, Acquisition, and IP Trends
- Cadence DAC 2015 and Denali Party Update