S3 Intelligent Tiering provides the ability to automatically move objects between storage tiers based on access patterns, without increasing operational overhead.

When Intelligent Tiering is enabled, you will be charged a small monitoring and automation fee on managed objects, however objects which are smaller than 128 KB are not eligible for auto-tiering and are therefore always stored in the Frequent Access Tier and will not incur the monitoring and automation fee.

How does Intelligent Tiering work?

By default Intelligent Tiering will move objects between 3 different tiers, based on how long it has been since each object was last accessed, which each support millisecond retrieval, 3 9s availability and 11 9s durability, although you can also opt to enable two additional archival storage tiers if your application can support slower retrieval times.

Let's quickly discuss the three default tiers;

First you have Frequent Access, which is the tier that all objects which are uploaded or transition to Intelligent Tiering will initially be placed in. If the objects in Frequent Access haven't been accessed for 30 consecutive days then they'll be transitioned to Infrequent Access which has lower storage fees. Further to this, if the object hasn't been accessed for 90 consecutive days then it will be moved to Archive Instant Access which will further reduce the storage costs while still providing millisecond retrieval times.

Now if you can afford longer retrieval times and have data which is rarely accessed you can enable one or both of the following archival tiers;

First up we have Archive Access which is a direct replacement tier for Archive Instant Access, whereby after the object is not accessed for 90 consecutive days it will be transitioned to Archive Access instead of Archive Instant Access thereby further decreasing storage but has the downside of minute-to-hour long retrieval times.

The final tier you can enable is Deep Archive Access, where after 180 consecutive days of an object not being accessed it will be placed within this tier, which is the lowest cost tier supported by Intelligent Tiering but comes with multi-hour retrieval times.

Conclusion

In conclusion Intelligent Tiering is an easy option to reduce storage cost if you have a large volume of data with unknown or unpredictable access patterns.