Introduction to Amazon S3

S3: (Simple Storage Service) is a secure, durable and highly scalable object storage service from aws.

It is an object-based storage: ie, it stores images video etc and not suitable for databases or other programmes.

Allows files of size from 0b to 5tb.


In S3, Files stored in Bucket. (a folder in the cloud)

Each bucket is a universal namespace ie. Each bucket require a unique name to identify itself over the internet

Bucket will sent a http 200 code to confirm successful upload

Data consistency model for S3

It is basically 2 things


  • Read after write consistency for PUTS of new objects.


Ie. we will be able to read the file(we have uploaded) immediately after the file got written(upload completed)


  • Eventual consistency for overwriting and deletes


Ie, if we change a file eg me.txt is renamed to me!!.txt we will immediately be able to read that file(me!!.txt) Note: delete or update might require a few mins, as they get stored in the edge locations

S3 is a simple key-value store

S3 is object-based file storage system, where object composed of the following things

  • Key (The name of the object). Eg. OR me.jpg
  • Value (Data) yea, the data made up of a sequence of bytes
  • Version ID (for versioning)
  • Metadata (data about data, which we are storing) eg: time when it was uploaded
    • Access control lists (where individual access on files)
    • Torrent (if we want to torrent this file)

Storage tiers and classes

S3 standard:

  • Built for 99.99% availability is guaranteed for S3 buckets (in SLA)
  • 99.9999999999%  (11 9s) Durability of information (ie this much of file will not be lost)
  • It has an ability to sustain in an event of losing 2 facilities concurrently

S3 -IA (Infrequently accessed):

  • For data accessed less frequently, but require rapid access
  • Lower charges, but a retrieval fee is charged
  • Stored in different zones

S3 – One zone IA (Infrequently accessed):

  • Lower cost of S3 -IA but stored only in one availability zone


  • Very cheap but for archival purpose only
  • Three types available – Expedited std and bulk
  • Std retrieval time is 3-5 hours

S3 Charges (pricing)

Charges based on the following

  • Per storage (GIG of space used)
  • Per request (Number of requests for the data, stored)
  • Storage management pricing (charged for metadata like access control)
  • Data transfer
  • Transfer acceleration, enable faster transfer of the file to a remote location between end user and bucket, using CloudFront’s edge location ref: image 1

Leave a Reply