From cwhite's wiki
- The name you assign your object
- A place to emulate a filesystem but in reality forms part of a key. S3 does not know about folders.
- Version ID
- Within a versioned bucket, the key and version ID combination uniquely identifies an object.
- The content you are storing
- A set of name-value pairs with which you can store information regarding this object. This is further broken down in User and System-defined user data. System user data will include things like: Creation date, storage class and server side encryption.
- Access Control Information
- S3 supports both the resource-based access control such as an ACL and bucket policies and user-based access control.
Bucket Name + Key = Object
- There is a soft limit of 100 buckets per account.
- S3 buckets are infinitely scalable when it comes to request and storage.
- A object can be up to 5TB in size with part sizes up to 5GB.
- An object can have up to 10,000 parts.
- Standard: General purpose
- Standard_IA: Infrequent access
- Glacier: Cold storage
- Reduced_Redundancy: Not as robust but not as expensive