CIO

INSIGHT: Why Kiwi businesses need to upgrade Windows Server 2003

“Server upgrades are more complex than end user operating system upgrades, so while you might have ‘got away with’ a reactive response to moving away from Windows XP, the risks are far too high for servers.”

For the New Zealand organisations who didn’t get the memo, or are simply operating in a bubble, Microsoft’s support for Windows Server 2003 ends on July 14 2015.

At present, Redmond estimates 23.8 million instances of Windows Server 2003 running across 11.9 million physical servers worldwide - astounding figures considering Kiwi businesses could still be running the about-to-be-obsolete operating system is less than 200 days time.

“Server upgrades are more complex than end user operating system upgrades, so while you might have ‘got away with’ a reactive response to moving away from Windows XP, the risks are far too high for servers,” says Brian Turner, Microsoft Practice Lead, Datacom.

“Server outages will at best grind your business to a halt, and at worst could cost you your entire business.”

Apart from the obvious business risk of running unsupported software, Turner says unfortunately the large installed base is likely to appeal to hackers, meaning malicious attacks will rise and be increasingly difficult to protect against.

For Turner, speaking as an ICT expert, there are two main challenges for organisations needing to migrate from Server 2003:

1. Knowing what they’ve got

With legacy infrastructure and changes in personnel over time, Turner says it can be “surprisingly hard” to identify what is actually underpinning operations.

“Remember too that the end of Server 2003 means the end of 32-bit applications,” he adds, “so not only do you have to identify all your affected servers, you may have to plan significant application upgrades at the same time.”

2. Knowing how to migrate

With major operating system upgrades undertaken only once every 4-5 years in many organisations, Turner accepts there may not be in-house expertise for a project of this magnitude.

“And you can bet that there’ll be a shortage of that expertise in the market the closer July 2015 comes,” he claims.

Overwhelmed?

But truth be told, many organisations across New Zealand, and the world, now understand the problem, but according to Turner, are completely overwhelmed about how to start solving it.

Where to start? Finding out can be a twofold process, as Turner explains.

“Firstly, by inspection – identify your physical and virtual servers and flag those that are running Server 2003,” he says.

“On each of these, identify running applications across a suitable time period for your business, to ensure you catch those items that are run infrequently. 6-8 weeks is a good starting point, but you may need to lengthen this based on your own business knowledge.

“Secondly, by discovery – just as there are automated tools that can monitor how an application is installed and run, there are tools that achieve this at the platform level. The combination of techniques should give you a solid understanding of your starting position.”

Page Break

Turner says the next step is to identify which of the servers and applications are most business-critical, or present most risk.

“A quick ‘traffic light’ classification – red for critical, orange for important and green for everything else gives you a starting point for planning your transition program,” he advises.

Going further, Turner says there are many options for how and what to migrate servers and applications to, calling on organisations to understand the technology and the implications it has on business processes.

The options for companies initiating change are two-fold - upgrade servers to Windows Server 2008 or Windows Server 2012.

“The move to 2008 is a simple mechanism to ‘buy some time’, and is typically appropriate for low-risk applications that will migrate cleanly and which you will then work out what to do with over the coming 4-5 years,” Turner says.

“Server 2012 is a better strategy for your more business-critical applications – you may need more than one tranche of upgrades though – first, the combination of the most critical and easiest to upgrade software, followed by successively more difficult migrations.

Advising organisations to replace some applications, or move them to the cloud, Turner believes for businesses, this could be the perfect time to investigate the costs and benefits of alternatives to older applications.

Future-proof

“This isn’t a one-off problem,” Turner warns.

While the Server 2003 phase out is “particularly tricky” because of the 32- to 64-bit issue, among others, Turner believes that the phase out of operating systems is cyclical, and in a few years you’ll be looking at Server 2008 and other migrations.

“With some planning and investment now, you can establish a new environment that will ease future migrations as well as this one,” he advises.

“Take a closer look at Microsoft’s Cloud OS vision, which it starts to support with a variety of features in Windows Server 2012 R2.

“The vision is to deliver data centre abstraction, which opens up the opportunity to develop new, overarching processes that are based on a series of incremental improvements and upgrades rather than a single, periodic major migration exercise.”

And while organisations may already have the skills to manage a migration or can acquire them in the short term, the question still remains – is it the best use of time?

“Perhaps you can handle the mainstream applications that are simple to move, but what about those outlier applications?” Turner questions.

“The ones without documentation or where the original developer is long gone, and no one is really sure how it works anymore, but they’re still business critical.

“Or perhaps you have all the documentation, but the cost and complexity of moving make the business case harder to justify.”

For Turner, the combination of the complexity of the migration away from Server 2003 and the opportunity to re-architect business environments to provide easier ongoing management mean that when the time goes come to upgrade, it’s best not to go it alone.