Microsoft taps enterprise apps

AS GENERAL MANAGER of Microsoft's .Net Strategy Group, Charles Fitzgerald has a front seat perspective on how the company's tools, server, and client initiatives come together. In the wake of Microsoft's release of the beta version of its Web Services Development Kit, Fitzgerald met with InfoWorld Test Center Director Steve Gillmor and News Editor Mark Jones to discuss the decomposition of enterprise applications, Microsoft's upcoming CRM play, and the role of Microsoft Office as Web services consumer.

Free IT resource Hear how top CIOs turn change into a competitive advantage. Sponsored by HP Free IT resource Attend the SOA Executive Forum: Breaking SOA Bottlenecks SOAExecForum.com/may2007 Sponsored by InfoWorld

InfoWorld: As Web services standards and tools evolve, enterprises are looking to integrate productivity and back-office applications. Is Microsoft still determined to encourage them to consider the .Net infrastructure stack in isolation?

Fitzgerald: Our fundamental belief is that the application is expanding beyond the scope of single systems, single devices, single organizations. The question is: How do we start to make integration intrinsic to the application model and accomplish it in a very Internet-native way? There was a period when people were really pushing the idea of one size fits all. And that model flies in the face of reality. Companies tend to have multiple, different applications. It's very hard to find a suite that does everything in a first-class way. If you're an energy company, there may be some vertical applications that you need. You probably have some stuff that's homegrown. You did an acquisition, you bought a company, and too bad they don't run the same software as you. So you need to look at how you can make those applications almost plug-and-play, in terms of being able to stitch those different things together, implement your business processes, adapt and evolve your business processes.

We have seen this fundamental shift in the way people think of business applications, and the best way I can characterize it is you're moving away from a model [where] the business application tries to suck everything into a single application. The business applications that win are going to be the ones that reach out or go centrifugal. Instead of the business application as a black hole, the new challenge is, how well can you reach out and connect to other applications, other assets, other partners, other suppliers? It's a fundamental shift in the way people have thought about business apps and the way they architect them.

InfoWorld: How do you see these toolkits intersecting with that new model?

Fitzgerald: The toolkits are all about giving people the latest and greatest Web services capabilities. The [Web Services Development Kit] that we announced was about taking the latest and greatest Web services standards, things like WS-Attachments and WS-Security, and making them available to the Visual Studio developer. We have a great extensibility model around Visual Studio. We obviously have the core base on Web services standards there, and what these new toolkits do is [provide] a way for us to deliver the latest and greatest functionality very, very quickly in a way that developers can get access to those things and start taking advantage of them as quickly as possible.

One of the key things about our approach is that toolkit will be a real production product; you'll be able to get product support on it. You'll be able to call us up and the license will let you deploy it. And one of the things I'd encourage you guys to ask everybody who's delivering toolkits is, "Is this a real production product? Can you get product support on it? Does the license allow you to actually deploy it?" There are an amazing number of toolkits floating around where, gee, it's just sample code and you're not actually allowed to do anything with it, which frankly doesn't help people a whole lot.

InfoWorld: I'm still looking for the connection between these toolkits and the centrifugal apps model, as you call it.

Fitzgerald: Well, we're providing infrastructure that allows people to build out a new generation of either applications or solutions that reach into existing, legacy stuff and take advantage of whatever assets are out there. We want to make that as quick and as easy and as accessible to the broadest set of people as we can. We want to make the whole set of Web services capabilities very, very broadly accessible, make it available very, very quickly to as many people as want to take advantage of that stuff.

InfoWorld: Are you partnering with SAP and Oracle on helping them to make enterprise applications more open or malleable?

Fitzgerald: I think there are two issues. One is, from a technology perspective, we're certainly providing lots and lots of technology that allowed these guys to build great Web services and expose information in a way that customers can much more readily take advantage of their existing information and their existing applications. I think the challenge is, from a business perspective, are these guys really committed to exposing those things? Take an example like our forthcoming MS CRM application, which is admittedly aimed at the low end of the market. But it's exposing all of its data as a native set of Web services, those things can be consumed by other applications. They can be exposed over the firewall so that you can connect them up with partners and customers -- all done in a very native, very seamless Web services way. If you're an existing vendor, there are architecture and transition issues to getting to that kind of a model. There also are questions about if they really want to do that. And that's a place where you guys need to go ask those vendors and get a sense of what it is that they're really committed to doing or not doing.

InfoWorld: What other types of applications besides CRM do you see coming?

Fitzgerald: I am not focused on what Microsoft's building. I focus on our work with different ICs across the industry, so it's pretty much the gamut from your traditional ERP and CRM, to all the little variants -- the ERMs and the SCMs. One of the ways I think about this is you have big application guys and small application guys. The big guys, like the SAPs, tend to have a very broad suite of applications, and some of those things they may actually do well, some of those things may be just part of the offering. Then you have the little guys, and the challenge for the little guys, historically, has been how do they compete with SAP or whoever it is that they're trying to compete with across the board?

One of the things that's interesting about this [centrifugal] model, where you have the ability to break [applications] up into smaller chunks of functionality, is that we think some of the barriers to entry go down. If you're a small vendor who's got some great ideas for a specific vertical or what the application guys would call a module, it turns out to be much easier to bring your product to bear than it has been historically when you had to do everything in order to compete with the big boys across the board. So there's some interesting opportunities there in terms of new entrants and smaller players.

If you're a bigger guy, in this world where the customer very much wants to be able to mix and match, incorporate new functionality in, write their own stuff, and get all these pieces to work together and have the flexibility and the agility to adapt and evolve those things, suddenly what [the bigger players] want to do is be the context, the broad framework for how companies think about their applications. Part of the way they're going to do that is by making it very easy for other people to plug their functionality in. Companies like PeopleSoft and Siebel are probably further down that path of thinking about that kind of model where success is predicated on providing a business application ecosystem for a variety of people.

So we've been spending a lot of time with the different application vendors, thinking about the technology requirements for more of a best-of-breed, almost plug-and-play type model for getting [application modules] to work together. And a lot of our investments are in the underlying infrastructure. So how do we make it really, really easy for people to expose their capabilities as a set of Web services? Or how do we think about pulling all those assets together? The best example of this is the work we're doing around orchestration with BizTalk -- if I have an arbitrary set of applications that need to come together, how do I implement a business process across those things in a consistent fashion?

InfoWorld: Do you see BizTalk being integrated with Office or the operating system on some level?

Fitzgerald: It certainly will work really well with all those different pieces. There's another dynamic that we're also pushing, which is that people have all this wonderful information locked up inside their business systems, but it's very difficult to get the information out. Whether it's your CRM system [or] your ERP system, how do you get that information out to the people who actually need it and can use it to do their jobs? I think we showed some of this stuff a couple weeks ago when you guys were up, about really starting to think of Web services not just as a server-to-server application integration methodology, but also [as a means] to expose information programmatically in a way that you can unlock that information and have it flow out to the arbitrary employees who need that information to do their job.

InfoWorld: That raises questions about the front end, the user interface, and whether that will be Office or the Web browser. What do you think is the future of the front end?

Fitzgerald: We certainly think Office is going to be a very significant front end for that data. It's not enough to just give people access to information, you want to give them the tools that allow them to analyze and create and make decisions and collaborate, and Office happens to be a great set of tools for doing those things. And two, it's obviously very, very widely deployed. There are hundreds of millions of people who know and use Office, and if you can get them the information they need to do their jobs, and put it inside Office, that's pretty exciting. What we found [when we did with our own CRM system internally] was that we had implemented this silo of information and it was hard to get people to put information into the system, it was hard to get information out. And it had its own proprietary front end. So what we did was really two things with Web services. One, we looked at the complete set of systems that track our interactions with customers, and it turns out there are about eight major systems inside Microsoft for tracking whether you've talked to customer service, tracking whether you've talked to product support, tracking what the sales force is doing. And we were able to bring all those things together, application-to-application, using Web services, and we actually did that implementation in about 100 days, which is pretty staggering if you are familiar with the EAI-type methodology.

The other thing we did was we used those Web services to unlock the data so that I, sitting at my desktop, can [be working] in Office, either Outlook or Access or Excel, and get access to that information. So instead of having to go through some CRM front end, I can pull that into the applications that I know, that I already use, that I'm comfortable with. And we've actually seen utilization of the CRM system pick up fairly dramatically once we made the information available to people in that way. It's this idea of unlocking the data, from a business process standpoint, from an application integration standpoint, and maybe most importantly, unlocking it from the standpoint of the employee who needs access to that stuff.

InfoWorld: What's your reaction to the portal suppliers and the enterprise app vendors who claim the browser is better suited to be the front end to these integrated environments?

Fitzgerald: I think they're missing the boat in a lot of ways. We could have a discussion about how much the dot-com [collapse] was [due to] the inability to deliver real value through the browser. The problem with the browser is that it's basically a dumb terminal. It's great for accessing information, but it's very poor for manipulating and analyzing and all that stuff. Think of it -- you're going to look at a stock. Well, I can go to the Web, I can get lots of information about it. But the way I'm going to make decisions is I may want to put that stuff into a spreadsheet, I want to chart it, I want to analyze it, I want to do some calculations. The browser's good at accessing information, but it's not very good at giving people tools to do something with [that information]. And so I think that the portal category is going to see some fairly dramatic transformations as people start to think about how they get beyond that dumb terminal model. How do they give people a richer set of tools to not only access stuff, but do something with it, make interesting and relevant business decisions?

There are a bunch of other problems with the portal, relative to not just the user experience. The developer experience is also crummy. These things are hard, they're expensive to deploy, they're hard to update and adapt. If I have a group of people that I'm doing some work with and I want to create a collaborative workspace -- that should be very easy for me to do. I shouldn't have to call my IT guys, and they shouldn't have to say, "Oh sure, we'll get you a portal. We'll get a couple of bus loads of consultants in here and we'll install some big iron, and in two years we'll have something up and running for you." You want to be able to do that with mainstream development tools, or if you've got the stuff up and running you want to make it very easy to adapt and evolve. There's just a tremendous opportunity to make the tools by which people access information and do something powerful with it. And that's a place where Microsoft has a lot of expertise and a lot to bring to the table.

InfoWorld: In effect, what you're talking about is skipping the portal phase and essentially doing a portal under the hood?

Fitzgerald: I don't think I'd characterize it that way. We're certainly interested in the next generation of tools for access and collaboration and giving people tools to actually do something with that information. And today's portal model certainly falls far short of where we need to be. ... There's a broader road map where I think over the long term you're going to see a hybrid model of [collaboration] services, and we're big, big fans of peer services obviously. There also are times when having server backup makes sense. From a product perspective today, the pieces aren't all in place, but we certainly are making lots and lots of investment there. I just think it's a little premature to talk about some of the next wave of products.

InfoWorld: What is your perception of the marketing noise that's coming from the Java space about the portal architecture?

Fitzgerald: The portal architecture is ultimately a dumb terminal strategy. And if you're a dumb terminal guy, it makes a lot of sense. One of the questions I've been asking customers since the beginning of the year is to characterize for me [their] most screwed-up development project. And nine times out of 10 it's a portal project, and they tend to be screwed up from a development perspective. And from a business result standpoint, they tend to underperform. People don't get all the information that they're looking for, they certainly can't do anything with it. The current portal approach is broken relative to the user experience, the development experience, the deployment experience, and the business return. If you're an app server guy, it's a great way to go make a really big complex sell. But the business returns just aren't there.

InfoWorld: How does the Microsoft architecture solve, for example, the problem of the information that they're looking for not being available?

Fitzgerald: We're focused on a couple different things. There's the issue of finding and accessing the appropriate information. In some cases, that's not trying to bolt the separate portal architecture onto a set of systems, but it's actually exposing that information programmatically. I gave you an example [about] Alchemy. Using Web services, we could make it easier to reach into these different systems, get to the data you want. From the standpoint of the user experience, once you've got access to the data, we obviously have a ton of very powerful tools that do whatever it is that we want to do with that [information]. We've talked a lot about where we're going with Office as kind of a universal client for accessing information inside the enterprise, outside the enterprise. All the work we're doing around Web services to unlock that stuff ultimately can come down and benefit the end-user.

You look at where we're going from a portal perspective. Think about all the development frameworks around the portals, where they tended to start as an HTML presentation thing, and then they're kind of backing into EAI. That's a lot of surface area for the different portal vendors to try to cover. We have a much more consistent architecture there. The next version of SharePoint will be totally built on top of the .Net framework; it uses ASP .Net. So instead of having a very small number of people who know how to work with a particular portal technology, in most cases, it's actually a way to drive consulting businesses for some of these companies. [With SharePoint] you're talking about millions of developers being able to go in and very easily set up the portal, deliver functionality for it, using mainstream tools, mainstream infrastructure. And then probably most importantly [there's] the ability to adapt and evolve it over time. Because all these things need to be constantly adapting and evolving to keep up with the information requirements that people have in the business environment.

InfoWorld: You mentioned companies that are using portals as a way of driving consulting services. IBM is one company springs to mind.

Fitzgerald: Yep. And IBM does what? Forty-plus percent of their business is consultant services. For every dollar they do in software revenue, they do $18 downstream in services revenue.

InfoWorld: Why then are they partnering with you so aggressively on Web services standards if that's going to essentially, according to your projection, drive out the complexity and raise the speed of development?

Fitzgerald: There's a fascinating dynamic going on inside IBM, and it's basically, are they a technology company or are they a services company? It's popping up not just in Web services; you look at their investments in management, there's an interesting little cat fight going on between the guys who are trying to take management out of it altogether and the guys who get paid by the hour for doing management.