Every year surrounding re:Invent, AWS announces a range of features that come with cost optimization opportunities. This year, the major news for cost savings is the introduction of Database Savings Plans, a new way for AWS customers to reduce their database costs across multiple services.

Commitment-based discounts (Savings Plans and Reserved Instances) are one of the easiest ways to save on AWS because they require no infrastructure changes or engineering effort; they're simply a pricing plan. That’s evidenced by anonymized Vantage customer data, which shows that for our customers, over 50% of EC2 compute spend is covered through Savings Plans and Reserved Instances. Of that, a staggering 46% is Savings Plans, while only 6% comes from Reserved Instances, highlighting Savings Plans as the preferred option when available (we’ll go over why in the next section).

However, before this month, Reserved Instances were the only commitment-based option for database compute. Reserved Instances for databases are widely used to achieve discounts of up to 69% compared to On-Demand pricing. For Vantage users, Reserved Instances account for 38% of RDS compute spend.

Now, with Database Savings Plans, AWS customers can save up to 35% on eligible database instances. While savings of 35% may seem lower than 69%, when comparing some of the same instances and term lengths, the list prices are far more comparable, and Savings Plans offer greater flexibility than Reserved Instances provide.

What are AWS Savings Plans and How Do They Differ From Reserved Instances?

Savings Plans and Reserved Instances are two types of AWS commitment-based purchasing plans that trade discounted hourly pricing on instances in exchange for one- or three-year commitments. If you use more than your commitment in any given hour, the excess is billed at On Demand rates. If you use less, you still pay for the full commitment. Additional discounts for paying fully or partially upfront for these plans may apply.

Both plans are available for instances across several AWS services (e.g., EC2 and Lambda) and now databases (e.g., RDS and Redshift).

The big difference between Reserved Instances and Savings Plans is that Reserved Instances have way more rigidity. With Reserved Instances, you select the specific instance type, service, region, and other attributes upfront and are locked into those specifications for the entire term. (There is some flexibility for EC2 and RDS Reserved Instances to apply across different sizes within the same instance family, but you're still constrained to that family and region.)

If you want to upgrade to a new instance or no longer need that instance type for whatever reason, you still need to pay for the Reserved Instance commitment, along with the new instance type you've spun up. Savings Plans provide significantly more flexibility by allowing you to change instance families, sizes, operating systems, and even regions while maintaining your discounted rate.

There are four types of AWS Savings Plans:

• Compute Savings Plans: The most flexible option, applying to usage for usage EC2, Fargate, Lambda, and more across any instance family, size, region, or operating system
• EC2 Instance Savings Plans: Less flexible but offering higher discounts, applying to a specific instance family within a region (e.g., m5 instances in us-east-1)
• Amazon SageMaker Savings Plans: Commitments for SageMaker usage, with flexibility across instance family, region, and size
• Savings Plans for AWS Databases: The newest option, covering databases like RDS, Aurora, ElastiCache, and MemoryDB, with flexibility across database engines and instance types

Reserved Instances are also trickier to manage because you may end up with dozens or even hundreds of individual Reserved Instances to track and optimize, each with its own specific configurations. In contrast, Savings Plans consolidate into a single hourly commitment that automatically applies to your eligible usage to grant you the highest discount possible, dramatically simplifying management and reducing operational overhead.

For example, for Compute Savings Plans, you might commit to spending $10 per hour on compute for one year (8,760 hours = $87,600 total commitment). AWS will then automatically apply this $10/hour commitment to your running instances each hour, prioritizing the instances where the discount provides the greatest savings. If you're running a mix of m5.large, c5.xlarge, and r5.2xlarge instances, AWS dynamically determines which instances receive the Savings Plan rate to maximize your total discount, without requiring you to specify or modify anything as your infrastructure changes.

For the sake of thoroughness, it’s worth mentioning Convertible Reserved Instances, which were released in 2016 before Savings Plans to address the rigidity issues with Standard Reserved Instances. They’re applicable only for EC2 and allow you to exchange your reservation for a different instance family, operating system, or tenancy during the term, as long as the new reservation is of equal or greater value. However, they come at a lower discount rate than Standard Reserved Instances and even Compute Savings Plans (up to 54% vs 72% and 66% respectively), and are therefore not recommended.

Savings Plans are the recommended choice most of the time for the flexibility and ease of use. However, there are a couple of caveats: Reserved Instances can occasionally be slightly cheaper than Savings Plans depending on the specific instance type and region, and zonal Reserved Instances guarantee capacity in a specific Availability Zone, whereas Savings Plans only provide the discount without any capacity reservation (though you can pair Savings Plans with On-Demand Capacity Reservations if needed).

Savings Plans for AWS Databases

AWS has been gradually moving away from Reserved Instances toward Savings Plans for quite some time, and the launch of Database Savings Plans is a long-awaited step in that direction.

Database Savings Plans provide a more flexible and predictable way than Reserved Instances to reduce costs across multiple database services. Customers can commit to a set hourly spend for eligible database usage and receive discounted pricing across supported services, including Amazon RDS, Aurora, DynamoDB, ElastiCache, DocumentDB, Neptune, Keyspaces, Timestream, and the Database Migration Service. The plans apply across different database engines, instance types, and deployment models, including both provisioned and serverless configurations (which were previously not possible to cover by Reserved Instances).

While the current discount range of 12% to 35% is lower than the maximum Reserved Instance discounts, the flexibility of Database Savings Plans allows customers to maintain discounted rates even if they change instance sizes, engines, or regions, which can result in comparable or even greater cost savings over time.

Some stipulations: currently, Database Savings Plans are offered for one-year terms with no upfront payment, simplifying adoption while AWS evaluates the best way to expand options. These plans cannot be combined with existing RDS Reserved Instances or DynamoDB reserved capacity for the same workload, ensuring that discounts apply predictably without overlap. Also, Database Savings Plans currently apply only to newer generation instances. If you're running older instance generations, you'll need to either continue using Reserved Instances or upgrade to newer instance types to take advantage of Database Savings Plans discounts.

Conclusion

This marks the start of a new era of cost savings for databases. Databases represent a substantial portion of overall AWS spend. RDS alone is the service with the second-highest share of Vantage spend at 11%, representing a significant opportunity for savings. For teams managing complex, evolving database architectures or undergoing cloud migrations, Database Savings Plans offer the cost predictability of Reserved Instances without the lock-in.