Thamarai's Blog

Some of the best things that happened to this company were because we shared. I'll give you some concrete examples. We opened the SPARC architecture and Cray ended up being bought by SGI, which had the Cray super server business based on SPARC. We bought that business for $20 million, which became our server business and that has got to be the best acquisition since DOS. And if we hadn't opened the community we couldn't have done all that R&D ourselves.

Same thing happened when we created the Java community. We spent a lot of money creating the Java Community Process and doing the church part as opposed to the state part. BEA, IBM, and a bunch of others went out and created Java app servers while we lagged in the market. The good news is that people didn't switch to .NET. They stayed in Java, and we were able to monetize other areas and come back later with our own app server, and we just reached 1.1 million subscribers to the Java Enterprise System based on our new app server and we priced it at model numbers that are way below because we had all of the community development to help us build that server.

Thirdly, we have now open sourced the implementation of the new UltraSPARC T1 Niagara microprocessor. How did we get this processor? We bought a little company called AFARA who was doing chip multithreading that used the SPARC architectures. It was binary compatible, it was a V9 architecture, we brought them in, we created an implementation that's blown away now, and now we are open sourcing the implementation and we've got dozens of universities moving their computer science departments to study the UltraSPARC microprocessor. We've got countries looking at standardizing on this microprocessor because it is open, zero barrier to entry, low-to-no barriers to exit. If you're sitting on a business model that is 40 percent after tax margins in software or silicon, this looks very scary.

We know how to make money and generate cash in this model because that is how this whole business model is geared up and we can get the best of the community, bring it on board, and staple it to our current price list with timely acquisitions. I don't think Sun could've survived our mistakes over the last 23 years without the community covering for us. Because when they leave and go to .NET, the barrier to exit is near-infinite, but when they stay within our SPARC/Solaris/Java community, we have a chance to come back later and make some money. So it is a very profitable thing and as I have committed over the last 23 years, until we get 10 times bigger than our nearest competitor, we will stay open. At that point, the economics of sharing suck.