As cloud cost optimization experts, we regularly get asked about how to optimize AWS Relational Database Service (RDS) costs. As part of our service, we analyze the RDS usage trends and provide recommendations for RDS commitments. However, our customers often also want to retain the same levels of flexibility that we deliver for our optimized compute services when they use RDS.
Issues with AWS RDS
RDS Reserved instances (RIs) are either 1 year or 3 year commitments. As discussed in a previous blog, we can rate commitments using three characteristics - discount level, technical flexibility and spend flexibility.
RDS RIs
| Discount: |
High |
| Technical flexibility: |
Low |
| Spend flexibility: |
Low |
While the RDS discount rate is good, the RIs have a very limited technical flexibility to change. Open source or Bring your own license (BYOL) database RIs are flexible in size but all RDS RIs are instance, database type and region specific. They also cannot be canceled, modified, or sold on the AWS RI marketplace, so have very little spend flexibility. There is commonly no 3 year nothing upfront option to commit, which further limits the long term commitment optionality.
Database Solutions
We have found that RDS is the correct solution for many situations including:
- If you have complex database support needs
- If you need license included databases for MS SQL or Oracle
- If you want AWS Aurora database for performance and resilient global service
However, if these scenarios do not apply and you have a reasonably straightforward set of database requirements, an open source or BYOL database, and you want a higher level of flexibility without paying on-demand prices, then you should consider 3rd party managed databases. Cloud Database core costs can be broken down into the following 3 components:
- The compute cost to run the database
- The packaging of the database instances
- The support of the database engine
AWS packages these costs into a single item and sells this as RDS, but it may be beneficial to unpack these costs. The compute cost can move from an RDS packaged cost to EC2, and by unbundling that cost, it opens up the options for a much more flexible approach as the EC2 compute optimization costs are much more flexible than the RDS RI costs. There are many providers of packaged databases on the AWS Marketplace for a huge range of databases, including the common ones available via RDS, which can be deployed on the EC2 instance of your choice. Leading providers include Cloud Img, Cognosys, Supported Images and Websoft9. Basic support is included in the packaging costs, and enhanced support is readily available via the packager of your choice. If you don’t require the enhanced support costs then you can easily save around $100 per month per instance, and even with an enhanced 3rd party support the database can still often be cheaper than using AWS RDS.
In summary, we recommend that you review if RDS is the right solution for you. You could save a lot of money by using a 3rd party packaged database and using Strategic Blue’s cost optimized services to provide great discounts and high levels of flexibility for your database.
To find out how Strategic Blue’s services can benefit you and your business, register here.