Now Generally Available: Autonomous Discount Management for AWS RDS, ElastiCache, MemoryDB, Redshift, and OpenSearch

We are excited to announce General Availability of Autonomous Discount Management (ADM) for AWS RDS, ElastiCache, MemoryDB, Redshift, and OpenSearch! After months of rigorous testing and product refinement, all based on feedback from our Early Access customers, we are now ready to bring these offers to a wider audience.

Our Journey from Early Access to General Availability

In November 2023, we introduced ADM in Early Access for AWS RDS, ElastiCache, Redshift, and OpenSearch. Our goal was clear – automate Reserved Instance (RI) management to optimize savings, mitigate risk, and save time for customers. We quickly expanded ADM to include MemoryDB in December 2023.

Over the past eight months, our platform has served dozens of Early Access customers of varying sizes with different workload patterns. To best address these real-world scenarios and after carefully analyzing the savings outcomes our platform produced, we made hundreds of product improvements. Here are a few examples:

We expanded the settings that govern our platform to better allow customers to control the behavior of our automation. For example, in addition to our default “Smart” mode which automatically selects match keys (region, instance type, platform, and size combination) for coverage, customers can now customize their strategy by indicating which specific match keys should be enabled or disabled.

We first developed Adaptive Laddering for AWS RDS, ElastiCache, MemoryDB, Redshift, and OpenSearch. We later applied this technology to ADM for Google Cloud Compute Engine which launched in February 2024, and Savings Plan Adaptive Laddering for AWS Compute which launched in May 2024. Each product learns from the workload patterns of its own Early Access users, and also from those of other products that leverage Adaptive Laddering. This accelerates the learning process. As our algorithms see more real-world scenarios, they are refined and become more battle-tested.

RDS, ElastiCache, MemoryDB, Redshift, and OpenSearch RIs are less flexible than commitment-based discount instruments for other services. With this in mind, we enhanced our Commitments dashboards with new metrics and a stronger emphasis on understanding and visualizing risk (more on this in an upcoming blog post).

Feedback from customers during Early Access has been resoundingly positive. They achieved greater savings (and risk-adjusted savings), surpassing what they could have done with DIY or alternative solutions. It’s not uncommon for mid- to large-scale customers to have dozens or even hundreds of different combinations of RIs to track and manage. Our platform automated away this time-consuming, error-prone process by managing hundreds of Adaptive Ladders concurrently. As such, our customers were able to focus on higher value FinOps activities that helped them and their businesses prosper (we love this as much as saving them more money).

We want to thank our Early Access customers. Their participation and feedback have been crucial in bringing these products to where they are today.

Key Features: Adaptive Laddering, Detailed Dashboards, and Control Settings

As we launch in General Availability, we are confident in our platform’s ability to broadly automate RI management for RDS, ElastiCache, MemoryDB, Redshift, and OpenSearch. Though there is more in store, here are the key features available today:

Automation via Adaptive Laddering^™ – When commitment-based discount instruments are rigid and inflexible, intelligently distributing commitment purchases over time can be used to safely increase coverage and derisk from overcommitment. This technique creates a staggered “ladder” where some portion of commitment continually expires. Based on changing usage patterns, the next commitment “rung” can either be renewed, increased, decreased, or expired off. Our algorithms watch dynamically changing usage patterns and intelligently adjust type, size, and purchase frequencies to create an optimal, ongoing ladder of RIs that adapts. A similar approach is used in personal investing to build bond or CD ladders with recurring liquidity events. With ProsperOps, this sophisticated investing strategy has been fully automated and applied to managing commitment-based discounts, so you can achieve exceptional outcomes, with little to no effort.

Detailed Outcome Visibility – A comprehensive set of dashboards are available to help you visualize and understand the rate optimization outcomes our platform delivers. Savings generated, Effective Savings Rate, utilization rate, spend coverage, and more are visible in the Savings dashboard. The Commitments dashboard of the console shows burn down, the shapes of Adaptive Ladders, lock-in risk, and remaining commitment obligations. In addition to global views, we now support detailed filtering so you can drill into outcomes for specific match keys.

Below is an example visualization of an Adaptive RI Ladder built for a specific RDS match key. Note that 1) multiple rungs are purchased monthly (our algorithms automatically adapt rung frequency), and 2) rung sizes vary as usage patterns change (our algorithms automatically adapt rung size). As the ladder fully matures, coverage will consistently be high, and an ongoing pattern of expirations creates flexibility that wouldn’t exist otherwise.

You are in Control – ProsperOps algorithms and automation execute on your behalf, but they only operate within the constraints you establish. Our platform has default settings, but you can also make changes to settings, such as minimum coverage, enabling/disabling certain match keys, and establishing vacate dates in anticipation of fully exiting a particular match key. Once settings are configured, our platform takes over the day-to-day activities to execute the defined strategy and deliver a savings outcome. Our platform always honors your settings and our Adaptive Laddering algorithms adjust accordingly.