Operating a public-cloud based infrastructure requires a careful appreciation of recurring costs and how to minimize them. We work with clients to optimize their cloud environments, and see several reoccurring themes when cloud cost optimization is done successfully.
Organizations that manage their cloud costs well deploy cost management tools and processes very early in the cloud implementation process. Your cost management tools should help you minimize costs by identifying:
Resources (storage, virtual machines) that were provisioned but have since become unused.
Oversized resources that were deployed with more CPU, memory, or storage than what the workload now requires.
Opportunities to make compute resource commitments and obtain discounts (AWS Reserved Instance or Azure Compute Pre-Purchase Program) with the optimal mix of commitment time, payment schedule, and flexibility.
Structure your cost management processes to be regular, inclusive, and automatic as possible. The cost management/optimization workflow should include a weekly review of the management tool's recommendations to determine the actual execution of the recommendations. Review the recommendations with applicable application and infrastructure owners to generate consensus and identify problem areas. Lastly, orchestrate and automate your changes to execute the recommendations as efficiently as possible.
Vendors may negotiate additional discounts for larger amounts of annual spend. Most cloud providers offer additional discounts above and beyond their compute resource commitment discounts for yearly spends that exceed $500,000. These discounts are typically more of a contract discussion than a technical review, but be sure to leverage the cloud vendor sales teams to help determine the optimal mix of technologies and discount programs to achieve the largest savings.
True long-term savings frequently comes from designing and migrating to cloud-native application designs and tools. Using cloud-native (“serverless”) platform offerings to minimize inventories of "always on" servers can yield large cost reductions but can require time and expense to fully implement.
We work with clients to optimize and embrace cloud-native application and infrastructure deployments. Let's discuss your cloud adoption challenges, comment below to start a discussion.