Longhorn and the Linux long-game

When the next version of Microsoft's operating system eventually launches, it will be into an environment unlike any faced by its predecessors.

Microsoft has said publicly that Beta 1 of Longhorn would arrive by the end of 2005, though internally, the company has been aiming for a release by the middle of this year. The final version of Longhorn is slated for the second half of 2006. But regardless of exactly when the launch finally happens, the threat from Linux on the server and client, and the quiet bleed to a thin computing model, has created some stiff challenges for the software giant, say industry watchers.

"We've seen a growing use of thin-client browser-based environments and this plays very strongly to the Linux desktop community. Longhorn is certainly Microsoft's attempt to reassert control against Linux and open source thin clients." says Neil Macehiter, a partner at MWD Advisors.

As a result, Microsoft is talking less about rich desktop applications, exploited locally on the client, and more about hybrid or 'smart' clients, which it is positioning as "the best of both worlds".

This is, according to Mark Quirk, head of technology at Microsoft's developer and platform group, because smart clients provide customers with the benefits of thin clients such as ease of deployment and administration, while also enabling them to undertake local processing.

"We see more people talking about smart clients than Linux. People want rich applications that they can deploy and manage centrally and you can do this with .Net today, but Longhorn will make it much easier," he says.

From a strategic point of view, however, it is actually the server rather than the client version of Longhorn that is more critical to Microsoft due to the long-term threat posed by Linux at the back-end.

While the majority of defections to Linux are currently taking place at the expense of Unix because migration is easier and the two can be looked after using similar management techniques, "this will change over time and Linux will seriously begin to threaten Microsoft sales as well as act as a replacement to it", says Macehiter.

"Given Microsoft's dominant position on the desktop, it's trying to make it play as well as possible with Windows server because it increases the barrier to taking out the server," he explains.

Olivier Nguyen van Tan, chief operating officer of Pierre Audoin Consultants, believes Microsoft will push its "integration innovation" message forcefully as equalling higher levels of productivity and lower cost for customers.

"Open source is a very big problem for Microsoft because it has seen open source and Linux coming in and winning market share. It can't fight on price because of value issues, the need for R&D investment and the fact that it has to maintain margins, so the only thing it can do is to provide another vision," he says.

The pitch from Microsoft will be that it can provide an integrated infrastructure solution where all of the individual elements such as the client and server operating systems, database, applications and the like will work closely together.

The software giant will try to position itself as a one-stop provider and highlight the fact that this level of integration in the software stack is simply not available in the Linux or open source world, although a raft of start-ups have emerged in the US over the last 12 months to try and address this very issue.

"Customers love open source because it's cheap and they think it's more reliable and secure than Microsoft, but you need different pieces and the problem is integration, which is real. This costs a lot of money so Microsoft is trying to fight it here and Longhorn is a big piece of that strategy," says consultant Nguyen van Tan.

He claims the only other player that can seriously rival Microsoft in its breadth of offerings is IBM, and Big Blue is basing its strategy for control of an organisation's infrastructure on Linux and Java.

IBM's message here is that its infrastructure framework is based on open standards that work in heterogeneous environments and which third parties can plug their offerings into.

As its infrastructure offerings conform to the concepts of portability and interoperability, IBM claims to be providing freedom of choice to customers that don't want to be forced down a pure Microsoft route or end up locked into the .Net development architecture and data model, of which Longhorn is a foundation stone.

"IBM and Microsoft are the two dominant industry players and it's going to be very much a battle royal to win mind share, particularly in the enterprise community. But it will be an interesting battle because they're both going in with very different weapons," says MWD's Machehiter.

IBM, on the one hand, comes from a position of strength on the server side. It has a reputation for trustworthiness and reliability in the data centre, a growing developer base and it is pushing a development framework that it is not under its control, but which supports heterogeneous systems and works well in an open source environment.

IBM is also starting to make a play on the desktop by promoting its Workplace strategy and particularly the front-end Websphere portal piece, which supports Microsoft's Office applications. The aim is to try and use the portal not only to control the user experience, but also as a Trojan horse to sell its backend infrastructure offerings.

Microsoft, on the other hand, is coming to the fight from a position of dominance on the desktop. It has a growing presence in the server space, a strong developer base and complete control of the development framework in the shape of .Net, which is geared to a Windows-based world — although again third party tools and applications can plug into it.

While Microsoft is always careful to fudge the portability message, it is now pushing the idea of interoperability strongly.

As Quirk points out, Bill Gates, the company's chairman and chief software architect, even sent out an executive letter to customers a couple of months ago indicating the importance of just this issue. "It's not a change of heart, although it might be a change in industry perception that we support interoperability."

"We've always had one of the most integrateable [sic] platforms, but the change has been in integration models. In the past, you had lots of middleware pieces gluing proprietary systems, but these days we use XML as a way of describing data. The old way of thinking of Windows as a proprietary system that only works with our products is no longer true," he claims.

But as Macehiter indicates: "The battle is not going to be won by locking customers in, in terms of interoperability protocols as in the bad old days with Corba and Dcom, so both parties want to co-operate here. They know that the battle will be won at the server and client level, so it's going to be an interesting one."

Van Tan, however, believes that neither vendor is likely to emerge as the dominant player. "It's not a question of who is going to win the war as neither of them are going to disappear in the next 10 days. The important question is what the balance between the two companies will be and we think it will be closer to 50-50 or maybe 45-55."

At the moment, he says, Microsoft with its .Net infrastructure "is still the challenger", but how this might change is very difficult to forecast into the future.

Nonetheless, there is a third — and currently more pressing — business reason as to why Longhorn is a crucial product for Microsoft, and that relates to revenues.

Macehiter explains: "One of the biggest challenges that Microsoft faces, and it's something that Gates has admitted, is the competition provided by its own installed base. Microsoft has to provide customers with compelling reasons to upgrade and that means increasingly getting them to sign up to the Software Assurance scheme because it provides it with predictable revenues."

Software Assurance is a software maintenance programme by which mainly enterprise customers pay a set annual fee and receive upgrades without having to pay extra as part of the deal.

The issue is, however, that, while such contracts last for three years, Microsoft released its last desktop version of Windows for the desktop, XP, in October 2001, and its last server product, Windows Server 2003, in April 2003.

As a result, points out Annette Jump, a principal analyst at Gartner: "Many of the support deals run out in 2006, so Microsoft can't really stretch the gap any more as it would start to have an effect."

While this is not so much an issue with small businesses, it an important consideration with mid-sized and large customers, she explains, because "they assume that they'll at least get something new. But if they don't feel they've had much value over the last three years, the danger is that they won't renew their enterprise agreements".

This feeling is particularly marked in Europe, she adds, which, over the last six to nine months, has led Microsoft to try and "add value through initiatives such as training and support to persuade companies to sign up again and in some cases, it's even been reducing prices".

Therefore, Microsoft will hit its 2006 deadline to ship the desktop version of Longhorn at any price, followed by the server version in 2007. According to Quirk, the client release is expected to appear in the second half of next year, and the server edition a year later.

But as Macehiter concludes: "Microsoft has been deliberately vague on what's in Longhorn and what's not. To hit its dates, it doesn't want to commit to functionality, and it will compromise, drop or reduce the scope of that functionality if it impacts on its ability to deliver. It's a very important issue for Microsoft and it's being very pragmatic about it."