We all know that cloud adoption is growing rapidly. Estimates from Gartner expect overall public cloud spend to reach $206.2 billion in 2019. Additionally, Infrastructure-as-a-Service is the fastest growing segment of public cloud and is expected to grow 27.6 percent when compared to 2018 spend. Typically, two-thirds of an organization’s cloud spend is on compute, which is especially vulnerable to waste.
With so many workloads being moved to the public cloud, wasted cloud spend is also up substantially. Gartner predicts cloud waste to hit over $14 billion in 2019. If your organization has invested in cloud resources, there’s a good chance some of that money is being wasted. This is money that could be put towards innovation, growth and development, rather than unused, idle, or overprovisioned cloud resources.
Idle Resources lead to Wasted Cloud Spending
One of the most common instances of wasted cloud spend is due to resources running when they’re not actively being used. This could be an instance where a development/testing environment is left running 24 hours a day when its only needed during business hours. It could also occur when a resource is no longer needed, and developers simply fail to terminate it.
Automations can easily be implemented to establish standards for resource run-times, thus eliminating the possibility of wasted cloud spend resulting from resources remaining active when they don’t need to be. Advanced Vector Automation (AVA) can automatically stop, terminate, or put on an off/on schedule for the following resources:
- Idle instances in Amazon EC2, Azure VMs, or Google Compute Engine,
- Idle load balancers
- Idle relational databases such as Amazon RDS, Azure SQL, or Google Cloud SQL
- Idle scale groups – auto scaling
Orphaned Resources Lead to Wasted Cloud Spending
Orphaned resources often take place when a VM is terminated, but the attached resources continue running, incurring costs.
Let’s use AWS EC2 as an example: All AWS EC2 instances in your environment have been stopped. However, you’re still getting monthly bills from Amazon EBS storage for unused instances. This occurs despite the fact that you stopped your EC2 instances from running because you’re still being charged for the amount of EBS storage originally provisioned. EC2 instances only incur charges while they’re actively running, but EBS volumes attached to those instances retain information and continue accruing costs even after the instance has been stopped. This can easily lead to wasted cloud spend.
Now you can automatically take a snapshot of the EBS volume as a backup before deleting the original volume. Snapshots are billed at a much more cost-effective rate and can be restored if needed later. Additionally, AVA can automatically back up EBS snapshots to S3 so they’re compressed, saving on storage costs.
Once you no longer need the snapshots, AVA can automatically delete the snapshots based on pre-established guidelines.
Over-Provisioned Resources Lead to Wasted Cloud Spending
Storage needs are always changing and need to constantly be right-sized. A recent study by RightScale found that 40 percent of cloud instances were sized at least one size larger than their workloads require. Typically, reducing an instance by one size results in a cost savings of 50 percent.
The 5 most common cases of overprovisioned resources that contribute to cloud waste are:
- Underutilized database warehouses
- Underutilized relational databases
- Underutilized instances/VMs
- Inefficient containerization
- Idle hosted caching tools
Rightsizing of resources is completely automated. You can establish threshold guidelines to ensure proper instance sizing is always performed and checked. This not only prevents wasted spend, it also saves valuable time spent provisioning and tiering whenever capacity needs might change.
The only way to ensure cost-optimized resources is through constant monitoring using analytics and metrics. AVA performs the tedious task for you by identifying and remediating improperly sized resources automatically.
Idle, orphaned, and overprovisioned resources will lead to wasted cloud spend exceeding $14 billion this calendar year. Some experts say this estimation is low.
Fortunately, cloud cost control is a major consideration for our customers. That’s why we at ASG have developed AVA, which applies machine learning to cloud computing analytics in order to implement proper sizing automations for cloud resources.