As general manager for platform strategies at Microsoft, Martin Taylor leads the software company's charge to contain and eventually eliminate open-source technology. Needless to say, that means Taylor does not shy away from controversy. In an interview with CRN Editor In Chief Michael Vizard, Taylor discusses the lessons Microsoft is learning from customer interest in open source and how the company may ultimately respond with more modular, component-based server offerings that would allow it create a more competitive solution in any given market.
CRN: On the face of it, one could conclude that interest in Linux is the market's way of telling Microsoft that Windows pricing needs to change. What message do you think the market is trying to send?
TAYLOR: I would actually look at a similar construct but a different answer. You have to ask one of two questions. Is it either a) Windows is priced too high, or b) are we offering the right product at the right price point? We position Windows server as a multifunction server that does a variety of things. So in some ways, we've got a McDonald's No. 5 super-size offering that costs $2.99 and someone just wants a Diet Coke that costs 99 cents. So do we cut the entire super-size No. 5 down to 98 cents, or do we try to find a way to just give somebody the Diet Coke if that's what they want?
From that perspective, there are a couple of interesting phenomena happening. As companies try to monetize Linux--i.e. Red Hat--the Red Hat enterprise server and their advanced server costs as much or more than Windows server, if you want it to be supported with security and everything else. If you model our security and our support and everything else, they actually cost more. It's not so much Linux monolithic vs. Microsoft. But it's Novell SUSE with one price point, one set of functionality; Red Hat enterprise server with one price point, one set of functionality; and Windows server. When you understand what you are paying for and where, then you see the relative price pressures come into play.
CRN: Sounds like there is a bit of soul searching going on. Where does all of this lead Microsoft?
TAYLOR: Here's the fundamental challenge, and I think we've solved it in one area. First, you have to segment the market into small, medium and large. Our [Windows] Small Business Server is the right product for the small-business market in the United States. Small-business customers only want one server, maybe two, and they might want a separate firewall. They don't want 10 servers all doing separate, desegregated workloads. Small Business Server was the first time that we really kind of architected from the ground up for the small-business environment. Then you have large enterprises on the other end. There are probably about 13 different server workloads that we have to think about there. The issue is, how do we offer a level of componentization to do those workloads and do it in a way that we don't lose the ability to make it easy, simple to manage and, when customers want to combine the workloads, not require them to combine two completely different products? You then have this [midsize]-business space that says maybe it's not one server, maybe it's not 13 and maybe it's four, where you have some app-type of server, some type of basic collaboration server. So that's the hard problem.
CRN: To make that work, you're then going to need some sort of built-in, federated intelligence that automatically discovers who's running what, when and where.
TAYLOR: Yes. It's not just as easy as taking out the existing HTTP server in Windows server and making it a separate product tomorrow. It is not that simple. It's bigger than, 'Do you have a pricing problem with Windows server?' quite honestly. Or for that matter, 'Do you have a pricing problem with Windows and Office on the desktop?' This is where you see an interesting dynamic playing out that asks what do people want to pay for, how do they want to pay for it, how do we extract value and how do our customers realize value in what we attempt to deliver? Those are some of the hardest problems that we work on in those scenarios. I feel good that that is the right way to address it because I think it benefits our customers and our partners by addressing it that way. We've got basically a couple of big server products. We don't have that level of modularity and componentization at the server level yet. We've taken a multifunction server approach. Our challenge is to provide a level of componentization without increasing management cost, without increasing complexity and without increasing security challenges--and all the while continuing to reduce cost and not losing the benefits you do get from the integration. It's a hard problem to try to figure that out.
CRN: Where do you see Linux being successful today?
TAYLOR: Definitely on the edge. You're just seeing edge services continue, such as firewall, appliances and those types of devices. Obviously, Unix migrations are happening. That's where, primarily, custom applications that people have written in-house are being moved over to Linux. But you're not seeing this huge ISV community created. Yes, some ISVs are being created, but not any massive ones. And the other place we see it is high-performance computing, scientific computing clusters that have lots and lots of servers.
CRN: Do you worry that Linux will gain more traction at the edge because of security concerns about Windows?
TAYLOR: Security is one of those workloads where Linux is getting traction, partly because we don't have a firewall appliance offering today. We have technologies, but we don't have a lockdown, hardened firewall that we can put in.
CRN: How do you respond to people who say that the open-source approach and the associated peer-review process inherently create better code?
TAYLOR: Just because you have more people looking at the code does not guarantee a level of quality, because those people might not be the most-qualified people to do code review. I'm not [making] a disparaging comment on the open-source community. I'm just simply saying that more in number does not mean it's more in quality. Let's just say that. That said, it's something that we continue to look at to see at what level and how do we open it up and share. And at the end of the day, there are only about 14 to 25 guys that actually check code into the Linux kernel. Just because you have a bunch folks out in the community that have the access to look at open-source product means that, by default, it will be more secure or higher quality.
CRN: What about Linux on the desktop? Is that a real competitive threat to Microsoft?
TAYLOR: We definitely see more conversations happening about Linux on the desktop in public-sector scenarios, primarily in emerging markets. A way to think about that is--back to our earlier discussion--there are Microsoft Office and Microsoft Windows in their current form and state. Is that the right set of products to deliver in an emerging market where you have a different set of challenges, from technology adoption, license pricing and usability scenarios?
CRN: Given the momentum of Linux, why do solution providers need to stay focused on Windows?
TAYLOR: By design, we've always moved out service and support from the core part of our pricing because we invest in the channel quite heavily. Our value-add is really in the R&D in the technology. So if you see price pressures come in on the services side of the model between the three companies, what ability do you have to sustain any type of economic model in which you don't end up going into a tailspin five, seven or 10 years down the road, if the only place where you extract value is in the services without the technology R&D? The opportunity is building, innovating, installing and customizing all of those things that our channel partners do on top of the R&D that Microsoft is delivering. As a channel partner, you've got to have two questions: Do you ride Microsoft's R&D wave, or do you ride this Red Hat Linux wave, knowing there's going to be some potential conflict with a vendor? The current value that the channel partner extracts from doing low-level integration will go away over time, and they'll have to move up. IBM and Novell have made acquisitions to try and get to an integrated platform that provides a level of functionality. Then the question becomes the broader ecosystem and the broader opportunity to do things to leverage that and the complexity associated with supporting Linux when issues come up with customers.