The revenues of the public cloud providers are currently $230 Bn and growing at 23% compound per year. It is estimated that of the paid-for capacity, at least 30% is not used at any one time. That is ~$80 billion of wasted spend.
So how do we go about identifying this wastage and removing it?
There are a range of tools out there, including tools provided by the cloud providers themselves. Some of these tools are better than others, some identify the size of the instance that should be running (Right-Sizing) whilst others identify the best way to buy it or run it and shut it down when it’s not needed any more (Right-Buying).
However, when the tools have identified potential savings in the cloud spend, the issue becomes ‘Do you want the tool to take the recommended action?’. The answer to this is often that companies don’t want automated cost reduction. And yet, the number of instances and the vast array of recommendations is too much for people to implement. When you ask them why they wouldn’t want software to be helping manage your cloud spend, they say ‘Just in case …’.
What they are concerned about is automated software shutting down an idle production instance which a while later is needed for a work-load. Obviously, where the workload is known and predictable, the automation can take account of that and not take action. Where it isn’t predictable, then leaving cloud instances running just in case it is needed in the future is a very expensive inaction.
Companies need to gain faith in the recommendations of cloud cost recommendation tools, and for those which are automated, allow the rule based implementation of the cost savings. Humans are never going to keep up with managing these vast estates.