Taking the Lead On Open Source

Steve Mills, a 31-year veteran of IBM, finds himself at an interesting crossroads in the company's long history. Obviously well-educated on IBM's proprietary software platforms and environments, Mills is charged with building bridges that will allow those proprietary technologies to cross over and operate in the new frontier of open source. Mills, a senior vice president and group executive at IBM Software since July 2000, sat down recently with VARBusiness senior editor Ed Scannell and outlined what IBM's general approach to open source has been, how he decides which products and technologies are best-suited for release to the open-source community, and the importance of standards.

VARBusiness: How big a bet has IBM placed on open source? Some have estimated up to $6 billion.

Mills: We made public statements about placing a $1 billion bet in early 2001, but that was not an annual statement. Let's put the investments in context. We spend a good deal of money on technology-enabling hardware, testing and making sure we are optimized to run as effectively as possible with the right distributions of Linux. We did our embedded processor on the zSeries [mainframes], as well as the native Linux support on the pSeries and iSeries platforms. I don't know if I can get a specific dollar amount.

VB: Do you think IBM, Novell, Red Hat and other key open-source suppliers can make a profitable living?

Mills: I think people have become comfortable that this is a viable model. When you think about the software business, it is not about just buying the software. Truth is, there is more money spent surrounding the software license than actually on the software license itself.

VB: Are VARs contributing to open source enough? Many are not as aggressive as they are in the proprietary markets.

Mills: In the past few years, we have seen significant growth in the business use of Linux, including many more midsize companies, so the opportunities for VARs have been on the rise. But VARs tend to not come at the market from the perspective of an operating system. They come at it with a particular class of buyer in mind. So chances are a VAR who already is carrying Unix in his kit bag will likely carry Linux.

VB: How do you know which products to release to the open-source community and which ones to hold back?

Mills: The thing about open source is, it gets traction and meaning if a community forms around it, but communities do not form around everything. We have lots of products in the IBM company, but we know that many of them would never be options for open source just because of the nature of the code, which might be too complex. It took us years to get Eclipse off the ground. We had to make a significant financial and technical contribution in the area. But open-sourcing pieces of code is not enough; you have to know how to participate in a community-based effort.

VB: In terms of standards, what is the next big area to address involving open source?

Mills: The more important work has to do with standards driven by customers--standards that deal with data definition and information interchange. These are profound in the way they affect the ecosystems inside those industries. We help by participating in a variety of industry verticals.

VB: What is your mindset these days for buying vs. building in terms of your open-source strategy?

Mills: I am focused on integration, period. Whether it is open source is really not the question. We have done six to eight acquisitions over the past five years, and they have all been middleware companies in various areas. We are looking for companies that can add value to the architectural model.