Zenasis Technologies is a five year old firm that brings transistor-level custom design optimization technology into an automated cell-based flow. Their main product is ZenTime which uses "Hybrid Optimization" to eliminate timing problems and improve the performance of cell-based designs. ZenTime eliminates the negative slack in cell-based design by exploring, identifying, designing, and inserting design-specific standard cells on the fly. The product integrates into most existing cell-based flows because the new cells use the same architecture as the user's basic standard cell library.
I had an opportunity to discuss the company with Dennis Harmon who came on board as President and CEO in July of 2005.
Would you give us a brief overview of your background?
My background is basically 20 plus years in design and design automation. It began back in the early eighties at Viewlogic. There were several other companies since then. Viewlogic was one of the larger ones who went public and subsequently was acquired by Synopsys. In the early 90s I and several others from Viewlogic founded a company called Synchronicity that was focused on deign collaboration tools for the semiconductor industry. That firm was acquired by MatrixOne in 2002. Now I am with Zenasis.
Would you give us a more expanded view of what Synchronicity was up to?
Synchronicity was basically building a product to help design engineers communicate and share information in real time over the Internet. What we built from a product perspective was the ability for companies with large geographically dispersed design teams to effectively work together in an environment where the location of the data was transparent to the users. For example, Intel was one of our largest customers and actually an investor in the company. They used Synchronicity technology and are not created by one group in one building. You'll get cores from different companies. You'll get IP from different companies. You will have design teams dispersed throughout the world working 24 hours a day on chip design. Our technology basically underlayed the EDA environment and provided a means for EDA data to be shared in an effective and efficient manner.
Who was Synchronicity's competition?
Strangely enough there was very little competition from a commercial perspective. The real competition for Synchronicity was internally developed solutions - people putting together scripts, PERL scripts, PDS and various other forms of configuration management tools. Of course the challenge was that it kind of fell apart as soon as you introduce an external partner, for example if you wanted to work with the ARM core team inside your design team. We were probably the only commercial solution and in fact Synchronicity is still the only commercial solution focused on semiconductor design. The uniqueness of that is that we understood all of the EDA data models, i.e. Cadence data base and Synopsys data base. We natively handled very complex data bases. There were some other small companies that did parts of what we did but not in the way we did it.
I am sure you know that the three major high end mechanical CAD vendors (Dassault Systemes with Catia, Unigraphics and PTC with Pro/E) generate significant revenue from PDM (Product Data Management) products and services. There are also MatrixOne and Agile who are focused on PDM and have revenues in excess of $100 million. How come there is no parallel in the EDA industry?
Laugh! That's a great question. It has always boggled our minds. If you think about it in the software industry, which we are in, we take seriously the challenge of managing product lifecycles and managing the process which is very interactive, very team intensive process of developing software. We have very sophisticated tools from companies like Rational Software which is a very large firm and now part of IBM that manages the information to control our software development process. With the mechanical CAD vendors as you mentioned there is the very same thing. Very large companies developing very sophisticated products manage the product lifecycle and the very interactive process of doing mechanical design through computer information tools. Why in the world the electronics industry has not grasped onto that was a confusion point to us. It provided us an opportunity to grown quite a large company. MatrixOne has taken that over and they are extending it from their traditional markets which were outside the semiconductor space into that space. I think that's good news for the semiconductor industry in general. More specifically it is good news for very large design teams that have to deal with this very thorny problem of managing data across their network. Lots of different people touching it. I have talked to many vice presidents and directors of engineering about chip tape outs that have failed not because of any sophisticated timing problem or sophisticated P&R problem but had failed simply because they used the wrong data. It's amazing!
Did Synchronicity seek out potential buyers like MatrixOne or did MatrixOne come a courting?
It was a combination. The way we perceived it, this was a very large problem to solve and for a small company to solve it we thought we would require a fairly large infrastructure. Right form the start of Synchronicity we believed that the way to scale the technology and the company was to find the right partner who had a global infrastructure to support this kind of problem of managing information across a global network. It's the biggest companies with the largest number of engineers that have the problem. In order to effectively support and scale the business we needed to rapidly find someone who could help us do that. So to that extent we sought out companies, in fact companies like Agile who was working on the periphery of the semiconductor space. They do data management of the electronics industry. They do have somewhat of a semiconductor focus but not a real business. And we obviously sought out MatrixOne as well.
How did you get comfortable sharing information with MatrixOne who might have ended up as a more direct competitor?
You know what's interesting is that the semiconductor design space is such a complicated space relative to the other spaces that they were focused on. From their perspective it was a build versus buy decision. If they went down the build road, they would have to gain a significant amount of expertise on semiconductor design, particularly the back end of the design. They couldn't acquire that in a reasonable time organically. As you now, it is very specialized space. We felt very comfortable with our core competency, the ability to have a high barrier to entry if they tried to do it themselves. In addition to that, I was very familiar with the executives over at MatrixOne. They knew what we did and we knew what they did. It worked out quite well.
Enough about the past. Would you give a brief overview of Zenasis, your current company?
Zenasis just turned five years old. It's a company focused on automating the design optimization phase of the semiconductor design flow, specifically standard cell approach to doing design. More specifically, the company is focused on reducing the time it takes design teams to meet the performance spec in effect in less time at lower cost. Before coming on board it was clear to me that they had a vast portfolio of very good technology. They had a great organization of people. I'm kind of quoting Geoffrey Moore's book the Crossing of the Chasm. They did a good job of getting to the chasm in terms of developing the technology and then developing a core group of initial customers and refining the technology. Now the challenge is to take that technology and turn it into whole products, not just the executables but the services, support, documentation and usability that's required to scale and deploy this type of technology across a large organization.
According to the press announcement you joined Zenasis last July. The founder and former CEO, Dr. Roy, moved over to CTO. This is not an uncommon occurrence in the EDA industry. What were your concerns coming into this situation and how has it worked out?