Autonomous Optimization for Cyclical Workloads Globally
I’m excited to announce our newest feature: Global Cyclical Optimization Support. Earlier this year, we launched a precursor capability called Advanced Cyclical Optimization Support. That update implemented cyclical workload optimization for EC2 usage at the region level. Specifically, if aggregate usage patterns for regional workloads using the same instance type, operating system, and tenancy formed a recurring cyclical pattern, our platform would maximize savings using advanced strategies to cover with commit discounts at the optimal point of the cycle instead of at the cyclical usage trough.
Even with region-level cyclical optimization however, it’s still possible (and in fact common) that total EC2, Lambda, and Fargate usage across all AWS regions forms an additional global recurring cyclical pattern. With the launch of Global Cyclical Optimization Support, we can now analyze aggregated global compute usage patterns and have extended the same algorithms we use at the region level to optimize commit discount coverage globally. Our platform decomposes real-time usage data into component parts, such as trend and seasonality, and combines with pricing and discount data to determine ongoing optimal coverage points. Our execution engine then acts to ensure your Compute Savings Plan coverage levels (which have a global scope) are optimal.
Maximizing Savings is Our Goal
Bottom line, this feature increases our customers’ Effective Savings Rate (ESR). Our platform is designed to maximize ESR, which is always correlated with maximum savings. Global Cyclical Optimization Support advances this objective by covering above usage valleys, where the total savings of additional commitment during usage peaks surpasses the losses incurred from overcommitment during usage valleys. The additional savings gained by covering above usage valleys outweighs the added costs up to a specific threshold, which is the point of max savings.
Below is an example of an actual 7-day global compute usage pattern (in purple) from one of our customers. Rather than cover at the trough (in red), our platform now determines the globally optimal max savings point (in yellow) and increases coverage to that point. In this example, the max savings point is 17% higher coverage than the trough, which leads to an 8% increase in savings.
When our platform has detected a global cyclical workload pattern and is covering at the max savings point, customers will see the following indicator next to their Savings Plan Utilization rate in the ProsperOps Console. Because we are now covering into the cycle, Compute Savings Plan utilization will be slightly lower, although savings and ESR will be maximized.
ProsperOps Console: Global Cyclical Optimization Support is Active
Talk to Us to Learn More
Global Cyclical Optimization Support is automatically enabled for all ProsperOps customers at no additional cost. Most customers with globally cyclical workloads will see an immediate increase in savings as a result of this launch. It is especially beneficial for customers capitalizing on the elastic nature of the cloud. Enjoy the extra savings!
Reach out today to learn how Global Cyclical Optimization Support and ProsperOps can enhance your cloud savings with less risk and less work.