In-Depth

Longhorn roundup

• By John K. Waters
• March 1, 2004

Longhorn, like many things in the history of Microsoft, is "long-awaited." A release date has not been announced yet, but it is still not too early to begin thinking seriously about Longhorn. "The thing I worry about with IT managers is that if they're looking too far down the road in terms of Microsoft's next release, they're going to take their eyes off what they should be doing now," said Michael Silver, vice president and research director at Gartner Inc., Stamford, Conn. "Don't sit around doing nothing waiting for Longhorn. Certainly understand what it's going to make you do, where you want to be and how you're going to get there, but don't ignore what's going on today, either."

In at least one way, Longhorn's long lead time is good for IT organizations, said Neil Macehiter, research director at London-based market research firm Ovum. Microsoft is providing unprecedented access to the code as it evolves, and both IT organizations and developers should take advantage of what is likely to be an ongoing education and re-education process.

Some industry watchers are saying that they would be surprised if Longhorn ships before 2008; certainly, we will not see it before 2006.

"What we will see certainly over the next year," Macehiter said, "is Microsoft building awareness of the capabilities of Longhorn. It needs to do this obviously to educate, but also to keep the momentum going because there is this delay [before the launch]. This is an opportunity for IT managers, third-party ISVs and application developers to really exploit that education process to understand what the implications of Longhorn are going to be for them."

In fact, Macehiter advises enterprises considering Longhorn to engage actively in a dialogue with Microsoft, to put the onus on the company as a supplier to explain what the new operating system will mean to their businesses.

However, if it is important for IT managers to begin to work out their Longhorn strategies, for developers it is absolutely essential to start scrutinizing early code and spending time on Microsoft's developer Web site and the Internet so that they are prepared to leverage the OS when it does ship.

"The big question right now is what should the development community be doing to get ready for Longhorn?" said Simon Yates, senior analyst at Forrester Research, Cambridge, Mass. "[The delayed release date] gives them more time to start figuring out exactly how to migrate their existing applications to Longhorn, how to build new applications that take advantage of Longhorn's features, and to make sure that the applications they use and are building take advantage of those features."

One of the most important issues from a developer perspective, he said, is the distinction between managed code in the Longhorn environment and unmanaged code, and creating interoperability between them. "All of the existing code, written pre-.NET basically, will fall into this unmanaged code category, and [developers are] going to have to do quite a lot of testing ahead of time to ensure interoperability," Yates explained. "Microsoft is trying to help out with that by developing things like the Common Language Runtime [CLR] and wrappers that allow COM components to talk to .NET components, but what it comes down to is if you have existing applications that you're going to migrate to Longhorn, start looking at what parts of the code within those apps are going to be affected and start thinking about how you're going to rewrite that code now."

But along with the advantages of the long lead time before the Longhorn release, said Gartner's Silver, there are some real disadvantages, especially for developers. "The advantage for ISVs of this longer delay until Longhorn is that you get more time to test. But it also means there's going to be a lot of change between now and the final release of the operating system -- one round of testing isn't going to do it. They're probably going to have to go through a couple of different development cycles with their applications to make sure they work in a Longhorn environment," he noted.

The key here for developers, said John West, principal consultant at Atlanta-based Microsoft consulting firm Intellinet, is to stay connected to those who are working on Longhorn, many of whom have set up blog sites. And Microsoft itself is soliciting unprecedented input from developers on the new OS.

"Microsoft is asking -- pleading -- for a lot of feedback from the user community on Longhorn," West said. "If you want to have a say in how some of these things work, there's a lot of opportunity right now to contribute." West recommends SharpReader.net and the RSS feeds as sources for developers.

"Historically, what has given Microsoft its tremendous strength is the way it has worked very closely with the ISV and developer community," added Ovum's Macehiter. "Microsoft has a developer community that is the envy of all of its competitors, and they're all trying to emulate it."

Don't skip XP
Longhorn's somewhat distant release date has created a much longer gap between major releases of Microsoft's operating systems, observed Forrester's Yates. Instead of a second edition of XP, which might have been expected, Microsoft has pushed back its Service Pack (SP2) release --originally planned for the end of 2003 -- until Q3/2004, and then nothing more until Longhorn. Fortunately, this will allow Microsoft to implement some of Longhorn's planned security and other features in SP2.

What this means for the enterprise, Yates said, is that IT managers have a substantially increased "preparation window." Essentially, the delay has pushed off any required investments in the new OS for another two to three years. But it also increases the importance of the SP2 update.

Meanwhile, it would be a mistake for organizations currently running older versions of Windows to wait for Longhorn to upgrade their systems, Yates warned. Plenty of organizations are running Windows 95 and 98 on the desktop, as Microsoft discovered when it announced last year that it was cutting off support for these systems. These organizations faced the risk of running an unsupported operating system, which meant that they would have to rely on their own IT department to respond to new security threats and to maintain and support those operating systems. Or they would have to find some outsourced service provider to take on the support burden. Within a couple of weeks of the announcement, Microsoft backed off, giving those companies more breathing room before support expires.

Now those customers are facing some tricky logistics, he said. "They have basically three choices. They can decide to stand pat with their existing hardware and OSs, and support them themselves after the Microsoft support expires. They can upgrade their operating system and run it on their old hardware. Or they can upgrade their hardware and operating system to either XP or Windows 2000 at the same time.

"Any hardware currently running Windows 95 or 98 is likely to be three to four or more years old," continued Yates. "The last big refresh cycle was ahead of Y2K, so those boxes are getting pretty long in the tooth. Windows XP sucks up a lot of memory, so it is likely to strain those systems, so there may be some hardware upgrades -- RAM at least -- that will be required anyway. We do know that Longhorn is probably going to require significant hardware upgrades when it does come out, so instead of having to buy all new hardware now and all new hardware again for the Longhorn upgrade, you can put off the hardware decision so that your acquisition costs are lower."

One alternative is to buy all new hardware bundled with new operating system software sooner rather than later. "That's a much higher acquisition cost, but at least in the near term -- say for the next year -- hardware prices are going to continue to decline," noted Yates. "Then as we start getting closer to the release of Longhorn, we'll start seeing those prices decline more slowly. There are a lot of good deals on hardware to be had over the next year, so 2004 is a good time to buy hardware if you're going to buy it."

Gartner's Silver agrees: "Questions come up on this issue: Do I deploy Windows XP or do I skip it? What are the odds that Microsoft will do another release between now and Longhorn? Based on today's understanding that Longhorn is the next release and when that's going to come, I'd say that you'll want to have some plan to integrate Windows XP, even if you're just bringing it in on new machines and moving to attrition.

"Think of it this way," he added, "if I stay on Windows 2000 and I have 100% of my machines running it in 2007, I'm starting to butt up against the end of support on that release and, even if Microsoft still supports it, it's a seven-year old release. Typically, application developers reduce their focus on an OS after five years or so. Consequently, if I'm still on Windows 2000 everywhere, what that may mean is that I'll feel pushed to move quicker to Longhorn and I won't be able to take that 18 months that I need to test my applications and make sure that everything is supported."

Look at the big picture
The key for IT managers is to consider Longhorn in the context of the organization. In other words, the basic Longhorn strategy should grow from a clear understanding of how the new OS supports the company's overall objectives as a business.

"Now is the time to start working out your strategy with Longhorn," said Ovum's Macehiter. "But rather than working out how you're going to use the particular capabilities within Longhorn, IT organizations need to very clearly understand what the real business benefit will be from exploiting those capabilities before getting their developers to look at it. It has to be done in the context of how it can help the organization to deliver real value to the business. And that is not an easy task."

Longhorn undoubtedly offers some significant potential benefits to businesses, he said. The key word here is potential. "It still has to make sense to the business," said Macehiter. "There's no getting away from it. There's no escaping the basic business need for this kind of assessment and analysis in the larger context, and managers ignore it at their peril, even when it comes to the advent of Longhorn."

In addition, IT organizations will want to consider Longhorn in the context of their supplier relationships. Enterprises must look not only internally at their own needs, but also at how their strategic partners are approaching Longhorn because those decisions have an impact on their strategy as an IT organization.

"I think the reality today is that organizations are increasingly looking to rationalize their supplier relationships," said Macehiter, "and Microsoft is clearly one of the players that will be considered in terms of reducing the number of suppliers. Microsoft is a significant supplier to a number of companies and will continue to be so."

Simply by adopting the new capabilities of Longhorn, said Macehiter, companies are sending the message that they are embracing Microsoft's vision going forward. "Look at Microsoft, IBM, Oracle, SAP and Sun, and ask yourself how those suppliers, going forward, fit with your IT organization," he said.

For San Jose, Calif.-based Adobe Systems Inc., one of Microsoft's biggest ISVs, planning ahead for new platform releases is an essential part of its ongoing management strategy. The company maintains a team dedicated to that purpose. Early access to Longhorn and a close working relationship with Microsoft actually allowed Adobe to create a prototype Longhorn-based application.

The Longhorn generation
Late last year, Bill Gates talked about trends that he believes are driving the next wave of software development, including Web services interoperability, the proliferation of smart clients, and various hardware and software innovations. In his crystal ball, Gates sees developers building connected systems using Web services "that reflect the dynamic nature of the real world"; software that will make "oceans of digital information more meaningful and actionable"; and new "breakthrough user experiences" that will improve interactions between people, groups and organizations.

"It's an exciting time to be a software developer," Gates said. "Continuing hardware advances, powerful tools and the potential of anything on the Internet to become a building block all make it a great time to do pioneering work."

It is a Microsoft-centric perspective, to be sure, but the software maker is supporting its vision of a "broad developer ecosystem" with a range of enhanced development tools and applications. And the technologies created for Longhorn are likely to reach deeply into these products. What might be thought of as the Longhorn generation includes a new version of Microsoft's Visual Studio development tools (code-named "Whidbey"), Microsoft SQL Server (code-named "Yukon") and some new smart-device software.

"I think enterprise users are caught in a difficult position at the moment," said Ovum's Macehiter. "Whether they like it or not, they are somewhat wedded to the Microsoft desktop environment today. I think going forward what we can see with Microsoft's strategy is that there will be a server version of Longhorn. We probably won't see the Avalon UI presentation services, but undoubtedly Indigo, the communication services, and WinFS, the storage services, will find their way onto the next iteration following Windows Server 2003. And that's going to impact application development."

What about choosing an alternative to Longhorn? Linux is enormously popular on the server side; how far could we be from a viable desktop Linux? And Sun is aggressively marketing -- and pricing -- its Sun Desktop System.

"They do have choices," said Macehiter. "Linux is the most obvious alternative on the server side. And they have J2EE as an alternative to .NET. With these choices, however, come significant migration efforts for the enterprise. It's about retraining users, and testing that all your applications that run on your desktop today will run on the Java desktop environment going forward. There is a choice, but there are pain points associated with that choice."

"Although Linux on the desktop is not something a lot of people are worried about right now, there are hardware manufacturers who are supporting Linux," added Forrester's Yates.

"New manufacturers are building tablet PCs on Linux. There are tablet PCs for less than a thousand dollars right now that run Linux. In the end, as a competitive strategy, Microsoft does have to try to beat back the wave of Linux and open-source software. Among their core base of enterprise customers it is a serious threat, and they are taking it seriously," added Yates.

"With Longhorn what I really think is that [Microsoft] is trying to reaffirm the Wintel platform and make it as easy as possible for developers to build applications for the Windows environment," he said. "And I think that they're trying their best to attract Java developers to their community by leveraging some of the development ideas that have been around in the Java world for a long time -- things like object orientation -- to try and grow their business by taking business away from those markets."

What about the problem of reluctant adopters within their own customer base? Yates believes Microsoft is expecting some reluctance and is planning for it accordingly. "I expect Microsoft to follow the same course: promote new features early, give developers plenty of time to understand how the apps they build work on the platform, discontinue support for older operating systems, monitor sentiment in the market for the new features and introduce with lots of fanfare," he said.

"I think this is a big challenge to the IT organization, and there is no simple answer," said Ovum's Macehiter. "What they must do is recognize it up front, understand what the costs and benefits are, as well as the challenges they'll face in choosing one over the other to deliver the same set of business objectives.

"There are no easy choices here," Macehiter added. "My word to IT managers is to drill into Longhorn -- not the technology per se, but what the technology promises to enable -- and then to match that with their business drivers and business imperatives. However inevitable Longhorn may seem to be, [IT managers] still need to make their decisions based on good business principles, what their goals are and the needs of their company."

Please see the following related stories:
"Zooming in on XAML" by John K. Waters
"The long and short of Longhorn" by John K. Waters

For more Longhorn article news, go to ADT's Longhorn Page