Overall Satisfaction with Amazon S3 (Simple Storage Service)
My current company is still in the process of moving to the AWS stack, but my previous company already had a successful story using Amazon tools, including S3. It was used across the whole organization. Server logs and data from the MySQL database were replicated to S3, and from there, the data got processed by Mapreduce programs running on EMR, and eventually get stored in RedShift or Elasticsearch. S3 is also used as a fast and neat backup storage for RedShift and Elasticsearch.
- Total number on the bills was definitely dropped when we moved to use AWS stacks from our own server clusters.
- "Always online" (nearly 100%) means less frustration dealing with power outages in your server plants and less people out there doing maintenance.
- DevOps has less work managing the user permissions, doing data replications, etc.
Amazon S3 is where you want to default to if you want to store a large amount of data. Compared to formatted data that you can store in Amazon RDS or DynamoDB, you can store your data in any format you want on S3. And the data retention policy can be really useful if you use S3 as a backup of your data elsewhere, or as a temporary location before data gets processed by other services. Compared to having your own servers, S3 provides a more reliable and maintainable service when your business grows.
Amazon S3 is best suited when you're already using an AWS stack or thinking about moving toward it. It is a reliable and high throughput storage that you will be using in the back of a lot of other Amazon tools such as EMR and RedShift. However, if what you need is a cheaper and a more longer-term storage, you might want to take a look at Amazon Glacier.