Microsoft sounds 100-day DOOM KLAXON for Windows Server 2003

well, let's see...

It runs 64-bit software nicely and it runs on 64-bit hardware well. Indeed, Windows 201x will almost certainly be slower on comparable hardware.

It don't need no new drivers. What it runs, it runs well and our consultant tells us that it's pretty much secure behind the firewalls and proxies he's set up.

Damn, new software and hardware is expensive. Tell me again why I need to splash out 30K for a new version of the server software (and the hardware, of course, required to run it).

Our technical person (read DBA-Support-Sysadmin-etc) has finally mastered Windows 2003. He doesn't really have the time to start on a whole new server. And he has exams coming up soon.

I performed my last 2003->2012 upgrade a while back. Three years? I thought I was being the stalwart on it.

Sorry, people, as much as you might say that you can continue with 2003 (of course you can, you could continue with DOS if you wanted to), there's a point at which it just doesn't make sense. When I transitioned it was hard to find 2003 drivers for RAID cards and things like that. I'm sure you have them.... for your current hardware.

When that RAID card breaks and you can't get a replacement (or you can, but a replacement server with RAID card is actually cheaper!), you'll struggle to get it across. And then you're into a 2012R2 install anyway, and virtualising what you have onto it. If you're bright enough to still have the full up-to-date system state to start from (Oh, you can rebuild the server you have with your data? Then on your head be it).

And then when you recreate your config on 2003, you'll find that much of it is obsolete and you could just use the built-in stuff in 2008/2012 to do what it does. In 2003 much of the useful stuff didn't really work well or at all. In later versions it's a core part of the OS.

I'm sure you can struggle on for another few years, with some IT guy that doesn't want to learn, with some users that don't want to transition, without spending a penny. And sometimes you can get away with that. When you can't, though, it's just embarrassing. I'm far from the "let's deploy cutting-edge just because" crowd, but there's a point at which the world has tested the software for you and written down all the quirks and workarounds and you should move forward.

As someone who's worked for schools with little money to spend on IT (including one that could not afford textbooks one year), done disaster recovery and freelance "consultancy" (yuk, hate the word) on places that are in dire need of IT help (and at least once said "I'm sorry, you need men on the ground each day, not consultants coming in and taking shots at what you haven't got, no charge for me telling you this"), I can tell you that IT is NOT a one-off expense. It never has been. If you put off the annual cost of it, you'll pay it in the future (plus interest!) in one lump sum, it's as simple as that.

And if you are large enough to have a permanent IT guy, then you're large enough to make him do his job (including professional development, future strategy for the IT, budgeting, costing and speccing out a new network every now and then), and large enough to afford an annual expense. Smaller than that, well, that's your business. But if you're running "servers" (plural) then there's really no excuse.

I'm sure 2003 will serve you well for years to come. And then the crunch point will cost you twice as much as it would have done to do regular maintenance. No different to never taking your car for service, not going to the dentist, or just leaving that leak dripping in your house because you "can't afford" a plumber.

Applications fail to work

The biggest challenge is that applications installed on windows cannot easily be extracted and migrated to newer platforms. Even if you have the original media it does not mean the configuration or data can be ported.

It's not laziness - it's a failure to constantly invest in the Software Development Life Cycle usually due to cost. App development - to make your apps work exactly the same but just on a newer o/s - is a bitter, but necessary pill to swallow.

Time to consider a Linux server?

If it's only plain old file and centralized email services plus data backup, like some small companies, maybe it's simply time to move to a Linux-based server?

To remain OK, just disconnect your business from t'interwebs. Simples.

best server since nt 4.0

2003 was the last version that had humans qa the product, and boy, does it show!

What should i "upgrade" to? Server tablet edition?

Lol, thanks but no thanks.

The answer, as always, is...

it depends.

There is a lot of milage in "if it aint broke, leave it the hell alone". But realize that it wont be getting patches for security vulnerabilities. If its an internal server that is isolated from untrusted code (esp. the internet), there may be no need to upgrade. The problem is that those guarantees of isolation are typically more imagined that real, and as others have mentioned, there will come a day (maybe in the distant future) where it becomes impractical to stay on an obsolete platform (need support for some new issue, need to replace hardware etc.) Your all grown ups, consider your personal circumstances and choose accordingly...