Microsoft's server chief talks cloud (Q&A)

It's been a busy year for Bob Muglia.

Microsoft's server and tools boss shipped an update to Windows Server, got promoted to division president, and prepared Microsoft's operating system in the clouds--Windows Azure--for its commercial launch.

In what has become a bit of a year-end ritual, Muglia sat down with CNET for a year-end interview. We hit on a range of topics, from the future of Windows Server, to why his bank won't be moving to Windows Azure any time soon, to the changing life of an IT manager, to Microsoft's consumer future. (Spoiler alert: Muglia thinks it is bright.)

Here's an edited transcript of our interview:

A few years out, how much does Windows Server, the server operating system, start to resemble Windows Azure?
Muglia: Well, making them as similar as possible is clearly the goal, and the goal is to take all the things that we do in Windows Server and make those capable to be done in Windows Azure, and then take the learning we have in Windows Azure and bring it back to Windows Server.

We just took the step of bringing the Windows Azure team, Amitabh (Srivastava) and his group, and putting that in my organization.

Now, what we also did as a part of that, is we merged the Windows Azure and Windows Server teams together. I just talked to Amitabh and he's really excited about the synergies that he can build across the organization and making these things as similar as possible.

In our own services, obviously we choose the hardware, and so there's a more limited set of things that work together. In some senses that gives us a bit of agility on the services side, because we can make something work in one very particular way, but what we've got a long term history of doing is understanding how to do that, and then abstracting that out to work in a much more general purpose way, to work with the hardware that our customers have.

One of the things we're looking at is how do we take the ideas that we're bringing to market in the form of Windows Azure service, and then build those into Windows Server and make them available to our on-premises customers and our hosters.

Do you expect there to be sort of an interim option? I wouldn't be surprised to see you guys do something in the intermediate term where you have the traditional Windows Server that runs on any hardware, you have Azure very customized for your data center, but then make a version of Windows Azure that they can write Azure apps for, but run in their own data centers.
Muglia: We're looking at those sorts of options. I think the trick to the thing is to understand what are the workloads that are most appropriate for doing that, and how would we structure that, and honestly we're still looking at that.

There are some interesting thoughts in that sense, and you can see how, for example, in a high-performance computing environment, where people could use hundreds or thousands of computers in one cluster, you know, Azure is really very, very helpful for something like that. But we're still looking at understanding exactly how we might bring some of those things to market.

I know you're the enterprise guy, but I thought I would give you the opportunity to come to the rescue of your consumer colleagues. One of the analysts pretty prominently said in The New York Times that it's kind of "game over" for Microsoft on the consumer side, particularly phones. I'm curious your thoughts on this.
Muglia: Well, I read your blog. You did, generally speaking, come to our defense. I mean, it's probably a fair thing to say that we understand the nerd. I think we understand more than the nerd, but it's certainly true that we do a good job with that audience, and that's actually my customer in a lot of senses, because I've got the developers and the IT pros.

You know, I feel like as a company we have a lot of focus on the consumer, and are doing a lot of great things that are quite revolutionary to consumers, and we're going to continue to do it. I mean, obviously if you look at what's happened even with Windows 7 and the success of Windows 7, most of the short term success has been in the consumer marketplace. The business marketplace is going to happen, but business moves slower than the consumer does.
Obviously we've had products like Xbox that have been very successful with consumer, and I think the new Zune work has been really fantastic, and obviously Bing has been really great.

I also think that people are going to be very pleasantly surprised to see the work we're doing in phones, and that will become visible next year. So, I think even in areas where there's been some concern, there is some really substantive, very, very innovative work coming down the line.

It seems a little weird to be saying this, but we're almost at 2010. Are there things in technology that you thought would have happened by now that haven't yet happened?
Muglia: Well, I certainly thought that the 787 would have flown by now. I hear it's supposed to fly Tuesday or Wednesday, though, so that's a good thing. I'm glad to hear that.

I think, if you had asked me in 2000, "Would we be further along on reading on devices than we are now?" I would have said we would be. And we're really starting to see that now with some very special purpose readers, but I would have thought it would have hit across more general purpose devices by now. So, that's probably been one thing.

You know, I think that that's related to how fast tablets or slates or whatever you might want to call them might have taken off. I might have thought that would have gone a little faster than it has. Those two are somewhat related with each other.

But those things always kind of come at different speeds, and I'm pretty confident both of them will become very important as we move forward.

If you could show the world as it is now to your 2000 self, what do you think would be most surprising?
Muglia: We've shifted to a world where all kinds of media are delivered digitally.

You think about what's happened with newspapers and magazines and things, I guess I wouldn't have predicted that shift would happen as dramatically. Not that it's delivered digitally, I think we would have expected that, but the kind of reporting and the real-time nature associated with it.

And I guess, similarly, some of the way social networks have developed, and the impact that they have had on the way people communicate. I find it so fascinating as an example to see the impact that Facebook has on what's happening in other countries like Iran, and getting information out, those sorts of things. You know, I'm not surprised by them, but I certainly wouldn't have predicted them back then.

So, probably your 2000 self might be surprised that in 2009 you're in touch with more people from your high school than you were in 2000?
Muglia: Right, exactly, things like that, exactly. That's a perfect way to say it.

Obviously, your focus is the IT world, and I'm curious, how different is the life of an IT manager? How much has changed in the last decade for what they do on a day-to-day basis?
Muglia: I actually think it's quite different for an IT manager. I think IT managers used to be expected to build the systems and do everything, and now I think they're much more focused on providing the infrastructure for the business teams and the people within business to do things.

One other thing that's missing at least from Microsoft at the end of this decade, as compared to the beginning is Bill Gates at least in a full-time sense. As one of the people kind of at the top of the technical ranks, I'm curious how have you noticed his absence in the last 18 months since he's left full-time work?
Muglia: I watched while Bill was here how his role shifted over a period of time. There was a period of time when the company was at a size and scope where Bill really was able to do the direction, the technical direction of very large parts of the company.

During the 1990s, we grew to a point where that was just not possible for a human being to do, although Bill's capacity is far beyond most. And so over time, Bill shifted into much more of an advisor role, and he provided advice and guidance.

While Bill's advice was always incredibly useful, he did a great job of also building a lot of people within the company that could also think in a similar sort of way to him. I mean, I'd like to think that I learned a lot from Bill and I'm able to do some of the things that Bill would have done.

I view part of my job is to take up and do many of the things that Bill did, and to do it in my area. My area of scope is broad but it's to a level that I think I can still do that effectively.

But I think what's happened is we now have a hierarchy of people. You've got Ray doing some significant cross-group things, and then you've got people like myself or Steven Sinofsky or folks in Stephen Elop's world, J Allard, all playing subset roles of what Bill used to do.

For a long time, Microsoft and you talked about this notion of autonomic computing. One of the things that conveyed was the sense that the IT would just sort of manage itself. Now you guys talk more about Dynamic IT. It seems like some of the idea that it's just going to magically happen has been moved from the notion.
Muglia: Let's just sort of kind of go back and talk about the time from where all those things sort of emerged in the 2002, 2003 sort of time. People sometimes called things autonomic computing, they sometimes called it utility computing. You know, our name for it was always Dynamic IT. It actually started out as the Dynamic Systems Initiative, and we sort of broadened it a bit with Dynamic IT a few years ago.

And the idea being that operational resources should be largely self-managing, and the process, the lifecycle of developing an application should be connected from the point of requirements, definition, through development, all the way to operations. That vision has not changed. We said it was a 10-year vision in 2003, and it will probably take us 10 years to really fully fulfill it. I think every year, as we release new products, we take substantive steps forward.

Now, the thing I didn't understand back then is that all of this would lead so naturally to the cloud application model, and that's what we've kind of put in place over the last year or two. I would very clearly say that is the next thing.

When you think about some of the biggest trends in the coming few years, I imagine cloud computing is a big one?
Muglia: Yeah. The thing that to me is so exciting about that is the impact I think it's going to have on business, most importantly allowing people to write applications more rapidly that really meet the needs of tier one enterprise apps, and do so at a fraction of the cost, both from a hardware perspective and from an operational perspective.

So, the kinds of things that would have required a mainframe in the past.
Muglia: Absolutely. Obviously we built software for very high-end systems. I mean, being straight and honest, that's not the main thing that our software is used for. We have a fairly small number of our systems running on these big $100,000 plus machines. But it's an important segment of the market, and clearly UNIX has a significant segment there, and the mainframe is a significant segment there.

When I talk to the Windows Azure guys, and I talk the software architecture that they've put in place in terms of having isolation units and understanding how to contain failures, and then I look at the way we are building the next generation data center systems, built inside these containers, I recognize that a container is essentially the mainframe of the future. The difference is that it's thousands of times more powerful than a mainframe at a fraction of the cost.

Do you have a sense of how many people are actually doing real cloud computing today, what percentage of businesses, and what that might look like in the next two or three years?
Muglia: I think Forrester has done some work that it's a really small number of people, like 4 percent of folks that are actually deploying things right now. So, it's still very nascent. But we expect a really large number of folks to start to (use cloud computing) over the next 12 to 18 months.

We're right at that inflection point where people are going to begin to start building real applications and begin deploying those applications into their environment, both for internal use and for their external customers. Certainly if you go out three to five years, we expect it to become very mainstream.

One of the constant debates is when people ask how much work will be done in a company's own data centers as compared to some sort of public cloud like Microsoft is running with Azure. Do you have a sense of where that mix might be a couple years from now or five years from now?
Muglia: I think certainly over the next five years we'll still see more work done in-house than in a public cloud. I mean, you'd have to move an awful lot of work out in order to shift that.

The question will be what is the cost and the effect, and at one level how much can this be done for people in a public cloud environment at a lower cost, and what level of security and trust can be established so that people feel comfortable moving their workloads to the cloud. I don't expect my bank will be moving their core financial systems to a public cloud environment in a five-year horizon, and that's probably a fine decision on their part.

I want to make, enable, and build the technology infrastructure to allow people to move their most sensitive data into the cloud so that some day it will become possible for a bank to do that, but I think it will take a little while for it to actually happen.

Last week, Microsoft bought a company called Opalis that specializes in software to manage data centers. Is this a big company? What made you interested in them?
Muglia: It's a moderate sized company. I'm really excited about this acquisition. I mean, what they do is something that's called "run book automation," and what they've done is they've built a very strong base of understanding of how to automate tasks that are happening within the data center.

And by the way, that's quite heterogeneous in its nature. Although they run fully on Windows, they're not limited and restricted to data center tasks that happen simply in Windows, but they can reach out and work with Linux and UNIX systems, et cetera.

Being able to automate a set of tasks is one of the key things that's going to be necessary to simplify the operations of any of these data center environments, and Opalis is a fantastic acquisition for us because they bring a ton of expertise and real world customer experience in that space. We think our customers will see value from this literally from day one.