Microsoft has been fuzzy on the expected ship date for the next major version of the Windows operating system, codenamed Longhorn. And of late, the company been raising the possibility of newly packaged editions of existing client and server products.
Bob Muglia, senior vice president of Microsoft's Windows Server division, confirmed that the server version of Longhorn will be released no earlier than 2006, and if a second release of Windows Server 2003 emerges, it will contain none of the "core" features of Longhorn.
Is there really going to be a Release 2 of Windows Server 2003?
There's a ton of innovation that we'd been building coincident with the release of (Windows Server) 2003, some of which has already shipped. We've shipped things like Windows SharePoint Services since (Windows Server) 2003 has gotten out.
And some things haven't shipped, like, for example, the Whidbey - Visual Studio 2005, I guess, is what I should call it now - the Common Language Runtime, those sorts of pieces of technology we know are coming in the next 12 months or so.
So we know there's a set of innovations that have been under development. We want to get those into customers' hands. Exactly the vehicle we're going to do that, we're still working on the specifics of that. We don't know all the details yet.
Do you think of Release 2 for Windows Server 2003 as an interim release before Longhorn?
I don't know how we'll package it.
When is Longhorn going to come out?
2006 is the earliest time frame we're looking at for Longhorn. The thing to realise about it though ... there's a lot of features coming in Longhorn. We want to make sure that those all fit together in a cohesive way and solve customer issues. We spent a ton of time over this last year or so on how we can make our existing systems more secure and robust. So all that benefit is going to transition into Longhorn as well. Teams are working pretty hard on all of those pieces.
Is it likely the client and server will ship at different times?
They will almost always ship at different times in the future. Clients need slightly less bake time than servers do. When we look at shipping a piece of software, we look at the process we go through for shipping it.
At the tail end of the process, we go through these long customer and internal deployment tests, which are measured in eight-week sorts of segments. So you deploy a version of the server out there, you leave it in production for eight weeks, you get all kinds of statistics about failures and things during that period, and then you start it again.
The clients don't have that same attribute. So there will almost always in circumstances be some distance between the times. Exactly how long, we still have to figure out.
Was the 2006 time frame applicable to both the Longhorn client and server?
I can't speak as much for the client pieces. I know the server piece.
Some analysts talk about Longhorn server in 2007. Is 2007 likely?
It's still too early to say for sure. I know we're tracking towards building this thing, and we're working very closely with the client group and the core group to drive it. We'll look at our tail-end schedules as we get closer, because to me, it will always be about the deployment side of it, making sure that the thing is solid and robust.
Is it possible that any Longhorn features will turn up in a second release of Windows Server 2003?
There are really three major pillars for Longhorn. One is the new user interface pieces in the Avalon UI, the graphical UI. Another is WinFS. And then the third is the web services infrastructure in Indigo. Those are three pretty big pieces, and I don't think you'll see any one of those three pieces.
At this point, it doesn't look like any of those things could possibly ship any earlier. Those are all about the platform transformation. That's really the core of Longhorn.
When I look at Windows Server, one of the things I do is look at what I call workloads that people use Windows Server for. So, for example, there's a file-and-print workload. There is an application server workload. There's a terminal server workload. There's a networking infrastructure workload. These are all different ways that people use Windows Server in their environment. Windows Server is used for many things, and these workloads describe it.
So what we're doing now, based on the (Windows Server) 2003 technology, is we have innovation going on in each and every one of those workloads. And some of that innovation may be ready for us to ship before Longhorn, and we'll try and get it out in whatever vehicle we can get it out.
Other pieces won't be ready, and they'll wind up in Longhorn. So in that sense, you could say, well, some pieces of technology will either ship in Longhorn or they might ship earlier. But the key innovations that you think about Longhorn, no, those won't ship earlier.
Carol Sliwa writes for Computerworld
This was first published in March 2004