Business Cloud News
Enterprises are massively over-provisioning cloud resources, under-using auto-scaling tools

Enterprises are massively over-provisioning cloud resources, under-using auto-scaling tools

A survey of 200 global CIOs published today suggests most businesses significantly over-provision cloud resources, with servers typically only seeing about 51 per cent utilisation over their full deployment cycle. Richard Davies, chief executive officer at ElasticHosts told BCN the results are a sign CIOs are bringing their legacy habits into the cloud world.

Traditionally, virtualised and bare metal cloud servers have been paid for according to the levels of capacity provisioned regardless of whether that capacity is being used or not and the independent CIO survey found businesses use just half (51%) of the capacity they provision on average.

The vast majority (90 per cent) of respondents said they see over-provisioning as a necessary evil in order to protect performance and ensure they can handle sudden spikes in demand. But contrary to what one would expect companies seem unwilling to pay extra for capacity used for just a small amount of the time, with 88 per cent of CIOs admitting they often sacrifice peak performance in order to keep costs down.

Juxtaposed with those results, the research found that almost half (49 per cent) of businesses believe that the cloud has delivered on the promise of being a “fully elastic solution.”

“Essentially, companies are paying for space they are not using half the time because they are running with extra headroom so they can handle peaks in performance. Yet at times of high demand, when arguably it is even more important that everything is working smoothly, web applications or websites will run slowly or even fail because they do not have the amount of server capacity needed,” said Richard Davies, chief executive officer of ElasticHosts.

“When enterprises were buying and setting up their own equipment this kind of over-provisioning was to be expected. A 2 to 1 ratio of over-provisioning is better than it was before – they have seen some benefits in the cloud for sure, but people are still carrying over their old habits,” he said.

Davies told BCN that this is partly related to how enterprises scale their applications. Just 14 per cent of businesses surveyed deploy fully automated cloud server infrastructure; the other 86 per cent are essentially running fixed sized instances and making occasional manual adjustments.

“There’s no shortage of cloud application automation tools out there but the research clearly shows most cloud users aren’t using these, instead opting for more manual, infrequent scaling,” he explained. “Part of the challenge has to do with the fact that those cloud auto-scaling solutions require you to architect your application in a way that works with these solutions, and they only work with classic 3GL web applications – HTTP servers and database servers. They scale the applications and don’t scale the databases, and they all assume you have an application layer written in a way that can it can scale horizontally.”

“If you went to a RightScale or similar vendor and said you wanted to have auto-scaling on fixed size VMs for my application, and my application was Oracle running on a Linux server, they wouldn’t be able to.”

The independent survey echoes the results of research produced earlier this summer which suggests cloud customers are frequently overcharged nearly £1bn annually, cumulatively, because of over-provisioning.