If you have asked yourself how can I reduce cloud computing costs, our guide to eliminating cloud waste should help you identify areas in which savings can be made. We not only aim to save your business money, but also time and concerns about governance and security. First, however, a quick explanation of how cloud waste occurs.
Cloud waste is usually a result of the over-enthusiastic deployment of resources in the cloud—both in terms of the number of resources and their configurations. Frequently, businesses—used to the sunk costs of on-premises IT environments—have added a little “wiggle room” in their provisioning to allow for future demand, and continued to pay for unused capacity when demand has failed to materialize.
The same psychology (the sunk costs of on-premises IT infrastructures) has led to resources being left running when they are not being used. It has been estimated that 44% of workloads are non-production workloads such as those used in quality assurance, development, testing, and staging, while a further 15% of cloud computing costs is accounted for by duplicated and no longer required resources.
Not all cloud waste is attributable to over-provisioning and leaving resources running when they are not being used. Some occurs due to decreases in demand, a build-up of old snapshots, or storage volumes failing to delete when an instance is terminated. All of these scenarios can be addressed quickly and easily using a cloud management platform to reduce cloud computing costs and optimize performance.
Rightsizing is the process of identifying incorrectly-provisioned resources and reconfiguring them to their optimum levels. The easiest way to do this is by allowing a cloud management platform to conduct a discovery of your cloud resources, analyze them, and make recommendations about their correct configurations based on usage data obtained from a monitoring service such as AWS CloudWatch.
The process also identifies duplicated and no longer required “zombie” resources, old snapshots, and unattached storage volume that can be terminated, as well as instances that could be placed into a less expensive pricing plan (for example AWS Reserved Instances) and data that is infrequently accessed that could be moved into IA storage or archived. It also identifies under-provisioned resources that can be rightsized upwards in order to improve their performance.
How much you will reduce cloud computing costs is subject to the volume of resources your business has deployed in the cloud and how they were configured on launch. If the estimates are correct that 55% of resources deployed in the public cloud are not correctly sized for their workloads, the savings could be significant.
Switching off resources used for non-production workloads is one of the best ways to reduce cloud computing costs, and there are several ways in which this can be achieved—manually through the cloud provider's portal, by writing scheduling scripts or by automating the scheduling process—usually one of the features on a cloud management platform.
Of the three ways, automating the scheduling process is the most cost-efficient. Developers can forget to manually switch resources off, and scheduling scripts can cost a lot to write and maintain depending on the number of resources you use for non-production workloads and the variety of schedules you want to apply, whereas the scheduling process can be automated with the click of a mouse.
On a typical schedule of “on” between 8:00 a.m. and 8:00 p.m. Monday to Friday, businesses save nearly 65% of the cost of deploying non-production workloads in the cloud. If your business works in irregular patterns or developers need to access their development instances at irregular times, automated scheduling software is sufficiently flexible that schedules can be quickly and easily interrupted.
Being able to reduce cloud computing costs is usually straightforward with the right tools. Maintaining the cost reduction often isn't. For this reason, many businesses are evaluating cloud management platforms with policy-driven automation. These platforms allow you to create rules (“policies”) about what actions should be taken when certain events occur. For example, policies can be created to:
Policy-driven automation saves you time keeping on top of your cloud cost reduction attempts and helps with governance and security too. Policies can be created that alert you to access control violations and suspicious activity such as log-ins from unrecognized IP addresses. In conclusion, there are a number of ways in which you can reduce cloud computing costs. Achieving these with the help of a cloud management platform has secondary benefits in relation to performance, governance and security.