Amazon S3 Overview

S3 stands for Simple Storage Service. S3 is a secure, durable, highly scalable object storage. 

S3 Feature Highlights

1. An S3 object consist of a key (a name), value (actual data), and metadata (version id and other data about your data).

2. Files in S3 are stored as simple key/value pairs in Buckets. An S3 bucket is synonymous to a folder. 

3. Object storages in general are flat file systems and does not have folders. However, S3 simulates folder hierarchy (like block storage) by allowing you to create folders within buckets. These are not actual folders, but the whole string with path and name, become the key to the file. E.g. myfolder1/myfolder2/myfile.txt.

4. S3 provides a simple web services interface through which we can store any amounts of data from anywhere on the internet. You access objects using the keys with a GET operation. You upload files using a PUT operation. If a file upload is successful, you get 200 OK HTTP status code.

5. S3 is a region level service. S3 service is not linked to any region, but buckets are linked to a region and you need to specify while creating the bucket. 

6. Amazon S3 is designed to sustain the concurrent loss of data in two facilities.

7. An S3 object support also following sub resources: Access Control Lists and Torrent (BitTorrent support).

8. S3 support tiered storage and has lifecycle management rules to decide what data sits on which tier.

9. Other popular S3 features include:

   1. Multipart Upload

   2. Encryption

   3. Versioning

   4. Tagging

   5. Server side logging

   6. Website redirect.

What is an Object Based Storage?

Object-based storage manages data as objects. Each object typically includes a key (a globally unique identifier), value (the data itself), and a variable amount of metadata.

Object storage has unique capabilities such as 

• access objects within an object storage using keys

• interfaces that can be directly programmable by the application,

• a namespace that can span multiple instances of physical hardware, and

• data management functions like data replication and data distribution at object-level granularity.

However, object storage is generally not suitable for installing an operating system.

Alternate storage architectures include file systems (e.g. hard disks), which are block storage, that manages hierarchical data as blocks within sectors and tracks.