What is Overprovisioning and How Does It Happen?

Determining the number of cloud resources that your organization needs is a balancing act. On the one hand, if you don’t budget for enough, you’ll find yourself going over budget in order to get the amount that you need. On the other hand, if you overprovision and invest in more resources, you’ll end up wasting money.

Overprovisioning is the act of deploying and paying for resources you don’t need in order to get enough of some other resource you want. In the cloud, the biggest source of overprovisioning is the way that cloud resources – like compute, memory, network bandwidth, storage capacity, and performance – are packaged together. If you need more of one of these things, you will need to pay for more of the others, regardless of whether you need them or not.

Why Organizations Might Overprovision

There are many reasons an organization might overprovision. One is to absorb the highest end-user demands. However, if the organization isn’t careful, it can be all too easy to forget to turn off extra resources when they aren’t needed – thus creating waste.

Another reason is to get faster performance. This pain is felt especially by organizations who are using the cloud for high-performing or mission-critical systems that require a high level of performance. In order to get it, they need to overprovision on other cloud resources, like capacity. Unfortunately, the cloud vendors set throttles on how much performance each of its customers can achieve since the cloud is a shared environment. Even with overprovisioning, cloud solutions are limited on their maximum capabilities and might not be able to meet the needs of all critical systems. All the money in the world may not necessarily get you the ultra-fast performance that you require.

The Pain of Overprovisioning

Overprovisioning is not unique to the cloud. Organizations with on-premises datacenters are used to overprovisioning on hardware in order to meet future growth plans and needs. However, since the cloud is self-service, on-demand, and virtually limitless in its scalability capabilities, it’s much easier to overprovision. And with the major cloud providers presenting pricing in a “Price per hour” list view, the costs seem so miniscule, many users don’t realize how much they have spent until the bill comes in.

Additionally, some cloud storage solutions require extra capabilities from the virtual machines (VMs) they are connected to in order to get the full performance out of the connection. This leads to overprovisioning of the VMs which increases the number of CPUs and, in turn, increases the cost of database software that is licensable by CPU cores, such as Oracle databases or Microsoft SQL Server. As such, you would need to buy additional database licenses for those additional cores (even though you don’t need the CPU power). To avoid paying for these extra licenses, you can choose to restrict the VM instance so only a subset of the total available cores is used. This is like logical partitioning which is used on-prem. This drops the database licensing requirement – and cost – but it also restricts the compute performance.

The snag? You still must pay the full cost of the VM, even though you are only using a fraction of its resources. So, while you’re avoiding unnecessary database licensing costs, you are still going to be billed monthly for compute cores that you literally cannot use.

Rightsizing Your Cloud Resources

If you’re currently overprovisioning your cloud resources, it’s a good idea to take a hard look at what resources you are currently using and how you can reduce them to fit your cloud needs more accurately. The biggest benefit of right-sizing your cloud resources is reducing wasted expenditure. But going through the process of rightsizing can offer you a better understanding of what your cloud needs actually are. How exactly are your cloud resources being used? Are you spinning up resources for Dev/Test purposes and not turning them off when you’re done – or are you spinning up multiple resources for the same purpose with no oversight? Are you using the best type of workloads and volumes for your needs? Right-sizing offers you better control over your cloud without compromising on price and performance.