Cloud not defeating on-premise as Windows Server 2003 prepares to hit support buffers

Martin Banks Profile picture for user mbanks March 31, 2015
As Microsoft’s Windows Server 2003 approaches the end of mainstream support from the company, creating a hackers’ target par excellence, is the image of cloud services as `disruptive technology’ scaring users away from a sensible solution?


One of the many features that cloud service providers boast of is the way that cloud is a `disruptive’ technology, taking businesses and individual users in dramatically new directions they could never believe possible before cloud services arrived.

The one trouble with that notion is that it is a): not half as disruptive as the hype suggests and, b): a message that is gloriously counter-productive as the last thing on God’s earth most businesses want is `disruption’, particularly to their core business processes. For them, disruption happens often enough without calling it down on themselves on purpose.

That could go some way towards providing an understanding of why the uptake of cloud services is still estimated as a low-teens percentage of the total IT business systems and software market. In a world of austerity and financial strictures, messing about with core IT resources is bound to be seen as a risk too far. Better to plough a more understood furrow, such as on-premise systems, even if the cost is higher.

July 14th this year sees this put to the test in a very real world manner, fort that is when Microsoft is withdrawing on-going support for Windows Server 2003, one of the great workhorses for businesses of all sizes, all round the world. Less than a year later – April 2016 – another Microsoft workhorse, SQL Server 2005, runs past the same watershed.

What is on-premise stays on-premise

At least users seem to be paying more attention to the implications of this more closely than they did with the withdrawal of support for Windows XP. According to a recent survey conducted by the Microsoft-centric Spiceworks online community, published earlier this year, some 15% of user have already fully migrated off Windows Server 2003, with another 48% already partially through the process. A further 28% were at least then at the planning stage. Only 8% professed to having no migration plans.

Set against this, a survey by AppZero estimates that 47% of Fortune 500 IT executives had no idea that Windows Server was so near the end of its support life.

More telling, however, are the Spiceworks findings that 56% plan to move to the latest version of Windows Server, 76% plan to move to virtualised systems, 30% plan to buy new physical servers, and 16% are planning to just upgrade Windows Server on the same hardware. All of that points to a strong desire to stay with some variant of an on-premise environment.

Support for that view comes with the finding that just 12% of those surveyed plan to migrate to cloud-delivered, or indeed any form of hosted services.

Estimates vary – and the number is obviously coming down – but in May last year HP reckoned there were up to 11 million Windows Server boxes running worldwide. The suggestion was that the migration rate would have to run at 25,000 units a day to get the job done. And because the trend is towards new on-premise installations, Spiceworks estimates that the total spent on businesses migrating is going to be of the order of $100 billion.

Microsoft itself will be charging users $600, per server, per year for Windows Server 2003 security support from July. It is not unreasonable to assume, if this current trend remains unchanged, that the same amount will need to be spent next year when SQL Server hits the formal support buffers. These are likely to be among the big revenue generators of the next few years.

The test for Microsoft

This is something of a test for Microsoft and its deeper, longer term cloud services aspirations. And it does suggest that the vendors – meaning not just Microsoft but also those members of its extensive partner community that have a vested interest in helping users move to the cloud – are not managing to get the message across.

The decoupling of business processes from specific hardware dependencies that comes with the cloud is an

Bar chart growth to cloud man on ladder © Warakorn -
increasing benefit to users. Applications can be run on any platform capable of running the right environment, and with business processes based on Windows Server that is just about any cloud service provider, on any datacentre anywhere in the world.

But it goes further, there is also the opportunity for users to decouple their business processes from the specifics of the application itself, particularly in terms of the Version Number, patches and Service Pack levels that are being used - updates that for many reasons are often forgotten or ignored by users. Depending on the level of service signed up for, the vast majority of these issues become the responsibility of the service provider.

Even if users have developed and built complex, dedicated applications on top of Windows Server, it becomes part of the service providers' remit to ensure continuity of service.

The Spiceworks survey indicates that many users have fallen into the classic trap of not keeping their Windows Server implementations up to date. They may well have installed all the security updates, but when it comes to Service Packs or even Version updates they have either let them slide by accident, or as a deliberate act.

This is often justified, especially in a time of financial austerity, by the fact that the additional features of a new version are deemed to be non-critical to the business, or because their adaptations to the application are such that moving them to a new version, with all the testing required, smacks of being far too disruptive to be worth the risk.

And given that many of those adaptations are likely to be mission critical to those businesses, disruption is the last thing any of them want to see.

This may be why so few of them seem to be keen on the idea of moving to the cloud, despite the fact that they now face the `brick versus hard-place’ decision of facing up to the disruption of change or risking becoming a prime target for the malicious hacker community, with all the damage that can be done by such people.

All of this does seem to add up to a lack of joined up thinking across the board, but within Microsoft in particular. Deferring the plan to can security support for Windows Server 2003 till at least the end of the year would have given a good overlap with the arrival of Windows 10. That would allow users to assess the potential for providing a real soup-to-nuts cloud-delivered environment that could give extensive operational management benefits as well a good platform on which to run – and develop – the latest current version of Windows Server.

Money in the sky.
It is also noticeable that many of Microsoft’s partner community have been really rather quiet on the subject, where one might expect to have drowned under an avalanche of `we-can-this-for-you-no-problem’ claims.

My Take

All in all, it does seem like one of the best possible arguments for businesses to move to the cloud is being allowed to slip through some collective fingers. It would be interesting to see a comparison of the estimated costs for a typical customer to migrate their Windows Server 2003 business applications to a cloud service, compared to the cost of buying new servers, new Windows Server versions, and engineering their adaptations and requirements to the new platform. A comparison of the relative stress levels would also be useful to have.

A grey colored placeholder image