Creating a Market

Gabe Moretti, EDA DesignLine, March 5th 2008

On Monday Imperas announced that it is providing some high-level processor models, APIs for building platform verification infrastructure and developing behavioral and processor models, and OVPSim, a reference simulator. These will be downloadable for free from an OVP website or from SourceForge, an open-source code repository.

The company has valued its donation to $4 million, which is not a small amount of money for a startup. My first question was if Simon Davidmann, the founder of Imperas, had suddenly decided to give his wealth away and join a monastery. But that illusion quickly dissipated, since I do know Simon, and although his debating abilities would certainly help any order he joins, the monastic life style just does not fit him.

So why would a startup decide to put all of its existing �products� into the public domain? The answer is simple: market creation. Although I have no doubt that what Imperas has managed to develop since its inception has significant engineering worth, the company has so far not met the high market visibility that Simon had predicted for his new company to anyone would listen shortly after he had sold Co-Design Automation to Synopsys. The market for Virtual Platforms is limited, and the competition from Mentor and Synopsys quite difficult to overcome because they can offer an entire suite of tools that carries a design team from architectural design through software integration and debug, to hardware design.

As the rest of the ESL companies can attest, this is not an exploding market, and growth is possible but slow and requires significant more investments than one could have planned a few years ago. Therefore, proving that he is not only a good engineer, but also an above average marketeer, Simon has instantly created a community, using the same market obstacles that would have challenged his company, to project it in the limelight and obtain support from the ESL community, always eager to find a new way to expand the market.

There are two major obstacles to the success of the Open Virtual Platform initiative. First of all, other companies must see the value of the site and donate valuable models and technology to it, and must do this is a short, say six months, period of time. Second, Imperas needs to build �for profit� products that will not conflict with free donated technology. In other words, its competitors can render its life quite hard, simply by donating to OVP technology that competes with what Imperas intends to sell.

The Bottom Line
The industry will welcome OVP and will try to use it, after all �Open-Source� is a key word lately since people equate it to �free stuff�. I think the following are the components of a successful OVP donation:

- The Imperas donations need to prove themselves valuable outside of the Imperas tools ecosystem.
- Processors and multi-processors vendors must provide free models to it.
- Other EDA companies must donate related technology to the site.
- Designers must find the site's contents valuable as building blocks for their own internal design flows. An open-source components that cannot be used, is worthless.

So let's wait. May be at DAC in June, you will find not only little placards in various exhibit booths proclaiming that the vendor is a member of OVP, but actual demos of how an OVP donation is being integrated in someone's flow.