AWS Storage Blog
Use cases for 7-day Amazon EFS Infrequent Access
Recently, Amazon EFS announced general availability of a new policy for Lifecycle Management, which moves files into the EFS Infrequent Access storage class (EFS IA) after 7 days of not being accessed. If you’ve been following along on the Storage blog, you may already know that tens of thousands of customers are storing petabytes of data on EFS at an effective price of just $0.08/GB-mo*. By doing so, they’re saving hundreds of thousands of dollars a year in the process.
*Pricing in the US East (N. Virginia) Region, and assumes 80% of storage is in EFS IA. See the Amazon EFS Pricing page for prices in other Regions.
So, when might you choose to use this new 7-day policy as opposed to any of the other available, longer duration policies? The obvious answer is to save more money, more quickly. Apart from the obvious, there are two common use cases for using the 7-day EFS IA policy – database backups and ETL.
We recently wrote about how EFS is great fit for database backup use cases since database applications expect file system storage and backup/restore processes are file-native. With the built-in data protection that EFS offers and the high availability/high durability that it’s designed for, you can be confident that your backups are ready to go when needed. EFS IA is a simple cost-optimization measure for DB backup, since all you need to do is essentially click a check box in the console to save money. You can minimize your recovery time objective (RTO) by keeping your backups natively in a file system. In other words, you don’t need to spend time or effort moving them between storage platforms. If and when it comes time to restore your database from backup, you simply run the restore command. With EFS and EFS IA, all your files are in the same file system namespace and readily accessible.
Our new 7-day policy is a great fit for DB backup use cases, since backups are far more likely to be restored within the first week of being taken (human error, system malfunction, etc.). By keeping newer backups on EFS Standard, you maximize performance of your restore operations. By aging off older backups to EFS IA, you easily save cost, remain compliant with your organization’s retention policies, and still achieve a low RTO.
Extract, Transform, Load (ETL)
We also see many customers using EFS for the storage aspect of their ETL pipelines. When building new ETL pipelines, or when adding or changing logic to an existing pipeline, sometimes customers will want to replay their ETL. They do this to test that it’s working functionally, or to handle a new edge case and ensure data quality. EFS Standard will optimize performance when replaying ETL on a source dataset, particularly when you need to iterate on it early in its lifecycle. Once you’ve settled on the business logic of the ETL job, the 7-day EFS IA policy will maximize cost savings, but allow for easily keeping extra and larger source datasets around for future testing purposes.
To get started with the new EFS IA 7-day policy, try it out in the EFS console.