Cloud waste has become a major problem for enterprises and estimated to be over $10B a year. If you have significant public cloud spend in your company it is probably affecting you. Read on to see some key tactics you can use to bring this problem under control, and proactively avoid the most common problems going forward.
1. Optimizing instance size
A common problem we see is oversized virtual machines, especially if end users are conservative in their decision making on capacity requirements. For first-time cloud users, the uncertainty in the cloud migration process often results in a straight mapping to the closest instance in the public cloud that provides currently provisioned resources. Where there isn’t an exact match, an instance that’s slightly bigger is picked. Analyzing the actual usage for VMs over a period can expose opportunities to right-size VM instances. Depending on the size of the environment, this can result in huge savings.
2. Selecting the right availability zone and region
While many customers pick regions based on needs for data proximity to their own location, they often forget that there are pricing differences between availability zones in a region. Another approach would be to consider alternate regions close to your current region of choice, which do not violate corporate policies or data sovereignty requirements. We have seen 10%+ variations in cost across regions. The savings need to be factored against the costs of migrating to another region.
HyperCloud Region Recommendations
3. Using reserved instances
Picking the right contract option (e.g. on-demand, one-year, three-year pre-commitments with partial or full upfront payments) can save substantial money. AWS estimate that reserved instances can save up to 75 percent on computing costs, while Azure estimates their own RIs can reduce costs up to 80 percent. The typical approach is to review recent usage patterns and reserve virtual machines in advance for long-running VMs or groups of similar instances that are expected to run continuously.
HyperCloud Reserved Instances Recommendations
4. Picking the right cloud
Not all applications need to be tied to a single cloud. This is particularly true for ephemeral workloads like those associated with dev/test. In such cases, you can take advantage of pricing differentials across AWS, Azure, and Google. Using a tool that can quickly provide you the costs across all clouds can make this a practical step in your workflow. Here’s an example from our platform showing cost results across AWS, Azure, and GCP for a workload.
HyperCloud Cross Cloud Comparison
5. Reduce wastage due to temporary workloads
Actively tracking VMs that are temporary allows you to set schedules to delete VMs and resources automatically. This, in turn, reduces the risks of dormant VMs and other resources adding significantly to cloud spend. These savings could also be achieved through setting policies for which cloud providers are made available to specific groups and users so that there is tighter ownership & hygiene.
6. Negotiate your enterprise agreement
Negotiating with your cloud provider can achieve reductions at scheduled rates. Understanding costs for your current instances across AWS, Azure, and Google is one of the inputs you can use to get a sense for discounts that you can request as part of your on-going negotiations and/or annual contract renewals with the large-scale cloud vendors.
7. Proactively deploying new apps to the right cloud
While most cloud cost reduction strategies are reactive and based on what has already happened, a proactive approach to cost management must be adopted to fully achieve potential savings. The ideal situation is deploying your applications to the right cloud, instance type and term from the beginning. Below is an example of the cost for an app, after looking at all major clouds and instance type to determine the best fit.
HyperCloud Cloud Recommendations
8. Migrate existing apps to the right cloud
Similarly, a migration is the perfect time to deploy your existing applications onto the right cloud, instance, and region. Completing a robust analysis of your application’s needs will help avoid unnecessary costs (and surprise bills) after the migration is complete.
Want to see your opportunity?
If you want to quickly see what cost problems may be lurking in your current cloud spend, sign up for our free 30-day trial here and see an analysis of your last bill and cost saving recommendations in under 10 minutes.
- 8 ways to save 50% on your cloud bill - July 17, 2018
- 10 ways MSPs can drive managed services growth in 2018 - January 5, 2018