IT chose to move to the cloud and cloud’s pay-per-use cost models because it wanted to operationalize instead of capitalize hardware and software. This has rendered hardware and software expenses discretionary instead of fixed, which potentially gives IT managers more flexibility to scale expenses upward or reduce them downward.
This sounds ideal, but cloud spending can also give CIOs and CFOs a false sense of security. Many believe they can turn their cloud costs off and on at will. With pay-per-use cloud, there is also a feeling that cloud resources are never wasted because you’re only paying for what you use.
But is this really the case?
When you run an application in the cloud, you're not only running the application, but also the underpinning data, network resources, infrastructure resources, storage, and security that are part of the application’s total workload.
Even if your staff has tools in the cloud that help manage your workloads, they don't have the same 360-degree visibility of resource utilization that they do in your own data center.
This cloud resource management problem amplifies exponentially when you add the myriad of cloud applications that end users bring to the cloud computing mix.
How, then, do you get on top of cloud spending as a major source of cost overruns in IT? Here are four ways:
1. Perform an independent cloud audit
In past practice, IT departments brought in independent cost auditors to look at telephony and data communications spend. This was helpful because the bills from telephony and data communications providers were so complex that IT couldn’t decipher them. Once the auditors broke down the bills and showed IT what it was spending, there almost always were opportunities to pare down costs.
Cloud computing is no different. The bills are complex, and this makes it difficult for IT to fully understand what it is getting for its money.
This is where an independent cloud cost audit can clarify the cost picture.
Once you have visibility of what you’re actually spending, you can work on cost modeling that more accurately captures what your IT workloads in the cloud require.
2. Think about the cloud like you think about your data center
An independent cloud cost audit will enable you to get your mind around what each application in the cloud is costing you to run. Just knowing this will get you back to the same feeling of cost control that you have in your internal data center.
The beauty of internal data center budgeting is always that you can fully track resource usage and spend of your IT resources. This enables you to calculate the cost for running the data center on an annual basis for purposes of budgeting.
With an independent cloud cost audit in hand, and information that would enable you to extrapolate resource consumption per cloud application, you can apply data center cost discipline in the cloud, even if you are using the cloud in a pay per use mode.
3. Negotiate cost discounts with your cloud providers
Once you know how much your annual cloud compute spend is, you can consider a more fixed spending plan with each cloud provider that is likely to net you cost discounts.
Like their customers, cloud providers like cost and revenue stability. If you can assess how much cloud resource use you will incur in a year and present this “known” usage to each cloud provider, you can negotiate a baseline fixed cost contract that each cloud provider will usually discount. You still have the flexibility of pay-per-use payments for anything that goes above these planned-for fixed costs.
4. Automate cloud resource use policies
Companies that leave their pay-per-use cloud consumption “wide open” for IT and end users will inevitably overspend. Sometimes overspending and resource upscaling are warranted, but there are also times when cloud resources that aren't being used are being charged for.
Cloud spend waste can be reduced if you automate your cloud usage policies, which you can do by using management tools that most cloud providers offer.
Here are several examples:
- You can automate the up-scaling of your computer and storage resources so that up-scaling only occurs at certain times of day, such as when you’re running peak loads.
- You can set job scheduling parameters in the cloud that automatically de-allocate certain resources when they are not planned for use (e.g., you can shut down certain cloud resources during non-business hours, when they aren't needed).
- Your cloud asset management system can identify resources that have been allocated in the cloud but that are no longer being used, and then automatically de-provision them.
As more companies develop their usage histories with the cloud, budgeting techniques will likewise improve.
The litmus test for IT leaders is whether you can sit down with the CFO or your staff and explain exactly what your cloud spend is, what you’re spending it on, and what that spend is likely to look like next year, or three years from now.
Most companies haven’t arrived at this point, but with the cloud resource management tools and cloud audit services that are emerging, there is every opportunity to improve cost performance in cloud computing.