S3 Features

S3 Lifecycle transition

• Following transition diagram represent how an object can be transitioned from one tier to another as per requirement.

• These movement of objects can be automated using lifecycle rules that can be attached to the bucket.

• Rules can be created per prefix or object tags.

Lifecycle rules actions

• Transition Actions

  • Allows configuring transitions to another storage.

• Expire Actions

  • Allows configuring objects to expire after sometime.

  • Can delete older versions

  • Can delete incomplete multipart uploads

S3 Analytics

• Helps identify when to transition objects to the right storage class.

• Offer analytics for Standard and Standard IA tier, does not work for OneZone-IA or Glacier.

• Report generated is in .csv format and will be updated daily.

• Starts generating analytics post 24 - 48 hours after enabling.

S3 Events Notifications

• Object creation, removal, restore etc actions have events attached to them.

• As all events goes to EventBridge, it allows 18 different AWS services as destination of the events.

• These events get delivered within seconds of the activity or sometimes even in a few minutes.

• The events need to be subscribed and appropritate IAM permissions i.e., Resource Access Policys should be attached/configured to the targets of the event.

S3 Baseline performance

• Automatically scales performance to high request rate with latency upto 100-200 ms.

• An application can achieve per-prefix in a bucket atleast,

  • 3500 PUT/DELETE/POST/COPY requests per second

  • 5500 GET/HEAD request per second.

• There are no limits to the number of prefixes in a bucket.

MultiPart upload

• Files with size above 100 MB are recommended, and with file size above 5GB must use multi part upload, as it parallelizes uploads.

Transfer acceleration

• Allows increasing transfer speed by forwarding the file to an edge location which will forward the data to S3 bucket in target region.

• It is compatible with multi-part upload.

S3 byte range Fetch

• Can parallelize fetches based on specific byte ranges.

S3 and Glacier Select

• Optimize data retrieval at server side by performing SQL query based filtering to retrive only required data.

Userdefined Metadata and S3 Object tags

Metadata

• If needed an abject can be assigned metadata which consist of some additional information about the object.

• Metadata are simple key-value pairs which are stored in lowercase.

• All the user defined metadata must have key prefix of x-amz-meta-

Tags

• Similar to meta data is also key value pairs

• Useful to segregate and assign finegrained permissions, analytics.

• Cannot search objects metadata or tags, just refer them, to search must have to integrate with other services like database or so to index based on tags or metadata as a solution.