Turbonomic Blog

AWS Simple Storage Service (S3)

Posted by Patrick Parish on Jan 27, 2017 11:28:32 AM

So the world’s going crazy over public cloud adoption and there is definitely credit to be given to Amazon for their market dominating and perhaps game changing, AWS solution. There’s some great infographics here and here featuring some powerful statistics on public cloud. Perhaps none more powerful is that 2016 marked the first year for Total Cloud Workloads to weigh in favour of public cloud at 56% with private cloud at 44%.

Amongst the AWS suite of cloud-based compute, network and storage solutions is their ‘Simple Storage System’ or ‘S3’. S3 is an object based storage solution and can be thought of as an easily accessed, easily managed and infinitely expandable data warehouse.

S3 Storage differs to Amazon’s Elastic File System or Elastic Block Storage that provides persistent block level storage volumes to use with EC2 (Amazon’s Elastic Cloud Compute) instances. S3 isn’t where you’ll ever install an operating system or mount a SQL database!

There are three tiers of S3 Storage available:

  • S3 Standard - Durable, immediately available suitable for frequently accessed data. By default, data stored in S3 is written across multiple devices in multiple locations providing resiliency. (SLA: 99.99% availability & 99.99999999999% durability)
  • S3 IA (Infrequently Accessed) – This is the same service as S3 although available at a lower cost. S3 IA users pay a retrieval fee meaning it is only a cost effective storage option for data that isn’t frequently accessed.
  • Reduced Redundancy Storage – A lower cost storage solution with reduced SLAs (SLA: 99.99% availability & durability)

Other AWS Storage Solutions?

In addition to S3 storage, AWS also offer Glacier Storage. Glacier is a low cost storage solution suitable for object based data that is not going to be accessed frequently or required with fast retrieval times as retrieving data from Glacier storage will take a number of hours!

Alternatively, Elastic Block Storage (EBS) is the standard storage mounted with every EC2 deployment. EBS serves only the EC2 instance they are built with and cannot be attached to other EC2 instances. Storage capacity automatically increases (hence ‘elastic’) and can be mounted to numerous EC2 instances. EBS offers the same storage service although can be mounted to numerous EC2 instances.

How does S3 Storage work?

Within the S3 service, users create ‘Buckets’. Buckets are used to store object based files and can be thought of as folders. When Buckets are created users specify which region the Bucket should be deployed in. Considerations here are usually cost, expected latency and any security or governance related policies.

Each object uploaded to an S3 bucket is independent in terms of its properties and associated permissions (who can and cannot access the file(s) for example). When individual or groups of files are uploaded to buckets, users specify the type of S3 storage to be used for those specific objects (RRS, IA or standard S3).

With Lifecycle Management S3 users can design lifecycle policies, automatically moving objects from one storage tier to another after a set number of days. These policies can be defined based on when the object was first created or when it was most recently accessed.

Why are some companies using it? What are the benefits?

Why are some companies NOT using it? What are their hesitations?

Where does S3 Storage fit in to your cloud strategy? We’d love to hear some great use-cases!

Further Reading:

Security with AWS and other Public Cloud offerings. - http://www.logicworks.net/blog/2016/02/amazon-aws-ecommerce-cloud-security/

Watch our On-Demand Webinar

AWS vs Azure - How to Unlock the Best Cloud Discounts

Topics: Cloud

Subscribe Here!

Recent Posts