Verification Cowboys

What does it take to be a successful EDA startup? Seven verification company executives provide some insight.

popularity

There was an event at DVCon that was both fun and serious. It was a panel of verification startup executives with the title “Ride with the Verify Seven.” Many of you know Raik Brinkmann, president and CEO of OneSpin Solutions who were the sponsors of the event, along with ESD Alliance, the organization that used to be known as EDAC.

The panel was moderated by Jim Hogan, a well-known venture capitalist for the EDA industry. He set the scene by describing the Valley of Death for a startup. “You sell a copy to one or two of your friends, but how do you turn that into a business? That is the valley of death. You realize you don’t have a product that you can see for everybody, so you have to build a better product and that requires funding. That moves you from angels and people you know to institutional investors, and these guys don’t care about technology or being good, nice people. All they care about is making money. If you don’t meet their expectations, you find yourself in the valley of death. It lasts until you can control your own destiny, and that is when you break even.”

Hogan had another important message for startups. “Timing is everything. If you have timing but you don’t execute you are dead. I want to move companies from being science projects to focus on getting product out and doing something that customer’s value. Value means they pay us.”

A growing number of people think that verification is becoming a limiter for the total cost of a chip and stretches time to market. This could mean that there is more potential for innovation in verification than other parts of the flow and seven verification companies were out to show that they were indeed going to avoid the Valley of Death.

One Spin’s Brinkman is quite the showman, or should we say cowboy. He dressed for the part and provided everyone with a few chuckles, as well as depicting what it can be like for the CEO of a small company to navigate through the early days of an EDA industry. It is not easy, as they discussed in this informal panel.

Andy Stein, vice president of North America sales for Avery Design Systems started with a cautionary note. “I see a lot of revolutionary ideas, but the key to getting out to market in EDA is they have to be done in an evolutionary fashion. Anything that is too different and doesn’t fit into the existing design flows is usually not going to make enough money soon enough to keep the company afloat.”

Rick Carlson, vice president of worldwide sales for Verific Design Automation has been involved with a lot of startups, many in the verification area. He observes that there are three reasons why companies fail. “One is they didn’t get the proper capitalization to start with. They spend too much money. Second is it takes longer to develop a viable product than they thought and third is takes longer to sell than they expected. Any one of them can kill you. But if you haven’t executed on the product, there is nothing that you can do to make up for that.”

Brinkmann suggests that startups may have an advantage over the more established players. “Startups today are very different from 10 years ago. It is no longer easy to find your key customer and grow it into a revenue stream. The technology is getting more difficult and you need more time to get a viable product out. There are several other challenges. One is the acceleration of the innovation speed. Our customers need to innovate at a higher pace, and they need new capabilities. But it is not always easy to find the right solution for the problem. That is also an opportunity for small startups because we can look into the problems deeper and quicker and provide solutions. That enables us to leverage a community of smaller companies.”

Prakash Narain, president and CEO for Real Intent, talked about some of the transformations they have made to follow opportunity. “Timing is when revelation meets opportunity. How do you prepare? Real Intent started as a formal verification tools company. There was an opportunity but the timing was off.” Narain then transformed his company into a static signoff tool vendor. He sees focus as being a key to success. “You go too slow and conservative. You will finish the race, but will not win. You go too aggressive and too fast and take too many risks, then you miss the gate and you do not win. The opportunity for startups is to focus on which gates you are going to take the customer through in the fastest possible manner.”

One area of verification that is receiving a lot of attention today is Portable Stimulus. There is no person better able to talk about this than Adnan Hamid, chief executive officer of Breker Verification Systems. Hamid described how the company had gone from concept to product through working with companies. Now he faces a different problem. “Along the way, it became a standards-oriented thing. Standards are a double-edged sword for a startup. On one hand, you are no longer on your own. But we are now trying to push what the standard should do because we understand the problem. The big guys are trying to steer the problem to something they understand.”

Phil Moorby, chief architect at Montana Design and the developer of the Verilog language, appeared to be frustrated with the state of verification. His criticism was in two areas. “Software simulation relies on the x86 architecture. x86 is good for some things, but it is not ideal for the verification process. The datapath architecture of the xX86 produces an inefficient cache behavior. Its multi-core direction has increased latency between threads in a parallel processing environment.”

Moorby was even more critical of another aspect of verification. “SystemVerilog testbench is the wrong language. The language has been a major duplication of effort and should have been C++ 11. C++ 11 was there in 2011 and SystemVerilog came out in 2012. I am not going to argue about how it should have been but we really have to be careful about spending a lot of time and effort on the language front. It is very subjective and almost like a religion.”

A very healthy discussion continued between the panelists and the audience, especially in the area of standards, openness, interfaces and collaboration. Brinkmann summed it up well: “Collaboration is important, and the ability to bring things together is important. You either have standards that you can build on, or you are in a bad situation. Our customers do not accept proprietary things because a startup may go out of business, or someone wants to shut them down. So they do not have the necessary safety. Industry standards are not in a good way right now because, unfortunately, the big companies are using that to slow things down, and that can make things hard.”