SAN FRANCISCO, Aug. 10, 2005 — Microsoft is not a name normally associated with the open source community, much less the community’s signature gathering, but attendees at the 2005 LinuxWorld Conference & Expo today heard Bill Hilf, open source industry veteran and former IBM Linux deployment specialist, talk about his experience running Microsoft’s Linux/OSS Lab.
Bill Hilf, Director of Platform Technology Strategy, Microsoft
While Microsoft’s presence at the conference may raise some eyebrows in the open source community — the last Microsoft executive to speak at the event jokingly wore a flack jacket — Hilf’s focus is on building bridges. Today, he led a session at the conference titled “Managing Linux in a Mixed Environment … at Microsoft?” To learn more about Microsoft’s position on OSS and the purpose of the company’s Linux/OSS Lab, PressPass spoke with Hilf, who is Microsoft’s director of platform technology strategy.
PressPass: What is your role at Microsoft and how did you come to assume it?
Hilf: Prior to joining Microsoft in January, 2004, I was a senior enterprise architect at IBM where I also helped lead the Linux technology strategy for a group that focused on emerging competitive markets. Microsoft had an interest in learning about and understanding OSS, including Linux, and brought me in as an expert in the field to lead a team and build a lab to do that type of work.
PressPass: What is the purpose of Microsoft’s Linux/Open Source Software Lab?
Hilf: It’s primarily an experiment in interoperability. By running Linux and a variety of other OSS in a highly Microsoft-centric IT environment, we’re learning how those technologies can better interoperate with Microsoft’s proprietary technologies. The lab consists of a few hundred servers plus a range of PCs, collectively running over 40 different Linux distributions, together with many different versions of UNIX. These various systems need to interoperate with the Windows-based networking, human-resources, e-mail and other systems that run Microsoft. In the process, we’re learning some very interesting things — like how to authenticate against Active Directory, how to run non-Microsoft mail clients with Microsoft Exchange Server, and so on.
My presentation at LinuxWorld was about how we manage the lab and the work we do specifically around interoperability. So many customers running mixed environments have asked how I manage Linux, UNIX and other OSS inside the world’s most Microsoft-centric IT environment. “How do you get these things to work together? How do you do deployment of software? How do you manage all of those different systems? What kind of hardware, software, and tools do you use?” Many people hold the perception that Microsoft is only about closed source and that we’re anti-open source. But the reality is not so black and white. Certainly, most customers don’t live in that either/or world. They choose a technology — an operating system or an application — based on its ability to solve a particular problem, not based on whether it’s proprietary or open source.
PressPass: What kinds of activities does the lab carry out?
Hilf: It’s a research lab, so there’s analysis, there’s testing, there’s benchmarking, there’s a variety of different interoperability scenarios that we work through, and there’s a large amount of resources that we provide to the rest of Microsoft so that they can understand and learn more about OSS. Essentially, we’re a center of competency for OSS inside Microsoft.
One of the biggest areas that my team and I look at that often doesn’t get captured is not just the technical analysis, but also the sociological elements of OSS and the community development model. We spend a tremendous amount of time understanding the community process of this model and learning how Microsoft can be more aware and its products more accessible to the community.
PressPass: What sort of reception have you gotten from the open source community since joining Microsoft?
Hilf: At a high level, it’s been extremely positive. I’ve attended every LinuxWorld conference that’s been held in the United States since it started in 1998, and many of the people who speak at these events are friends of mine that I maintain good working relationships with. Part of my job is to serve as a bridge to the open source development community, so that they have someone that they know and can communicate with at Microsoft.
PressPass: What do you say to those who speculate that Microsoft is considering porting to Linux or its own version of Linux?
Hilf: I’m often asked whether Microsoft is working on its own Linux implementation. The answer is, absolutely not. We are 100-percent committed to the Windows operating system and we see in it a fantastic potential for innovation – much more so than in Linux. The open source process has arguably turned out reliable point solutions at the edge, but the closer you get to the core of the application stack, the greater the complexity.
OSS proclaims a virtue of no design by design, and OSS programmers tend to solve interesting problems that are relevant to their individual domains. But these solutions often show deficiencies in integration. Without starting from an architectural point of view, the programmers’ inability to foresee integration points into an unknown problem domain – anything outside their own 9-to-5 experience – tends to result in a plug-in mentality, or a reliance on loosely coupled architectures. Some level of tested, architected integration capability has to be present in the platform to support enterprise-level applications, as well as consistent desktop experiences. Windows offers that capability in abundance.
PressPass: How is the work you do in the lab related to Microsoft’s interoperability commitment?
Hilf: I don’t own interoperability for all of Microsoft, but what I do provide for all of our different product groups is a central place — from a Linux/UNIX/OSS perspective — where I can test their products in a way to guarantee that they truly are interoperable.
There’s no better environment to test a UNIX interoperability capability for a Microsoft product than in my lab because I run dozens of different versions of UNIX and Linux. I have more mixed environments than most sane customers would ever have. I tell most of the product teams, “If you can make it through my lab, you’re probably going to survive 99.9 percent of the customer environments you go into.” Sometimes, I go and test the products on my own. Other times, a product group may come to me and say: “We’re hearing from customers who want to understand how these scenarios can possibly work. Can you help us?”
It’s worth noting that interoperability is not a new focus for Microsoft. We have been enabling interoperability via Host Integration Server and Services for UNIX and the like for some time, and these efforts continue. A key approach to bridging the interoperability divide at Microsoft is our strong support for open standards, as seen in our involvement in Web services, XML and SOAP. It is these open standards – not open source – that help make today’s integration technologies more interoperable than ever before, thereby minimizing vendor lock-in for customers. That said, we believe that incumbent vendors like Microsoft and vendors with emerging platforms need to share the responsibility of bridging the interoperability gaps and working together to meet customer satisfaction.
PressPass: What is Microsoft’s long-term strategy for the Linux/Open Source Software Lab, if any?
Hilf: Right now, my lab gives the company a line of sight that it really wants and needs into OSS. When I built this lab and this technical team, I had a long-term strategy in mind that involved a commitment to interoperability across all the product groups. That way, there wouldn’t need to be a centralized place for this kind of competency, because people would constantly be thinking about interoperability and management issues in these environments. Many product groups at Microsoft already do this, but I’d like to see it happening across the board.
It is a very interesting time now for Microsoft, particularly with all the new products we’re developing and the waves of innovation coming out with Windows Vista. There’s plenty of work for us to do in the near term, but eventually I’d like to see this testing and analysis as part of all Microsoft products’ DNA.
