Amazon S3 - Quick Thoughts

I’ve been trying out the Amazon S3 service over the past few weeks storing, listing and retrieving files. Here are my thoughts on the service.

Client API

I have been using the JetS3t API and found the API simple and easy to use. Though there is not much documentation, the samples provided are enough for a Java programmer to understand the methods. There are a few sample applications provided too. Notable among them is Cockpit, which comes with a GUI to perform operations on files stored in your S3 account. There is a good tutorial using the JetS3t API available here. Amazon has its own API too but I did not like it much as I found JetS3t simple and intuitive to use.

Service Design

The contents in S3 are stored as key-value pairs just like in a Java Hashtable. The key is called a “Bucket” and is the unique storage location in S3. The value is the content of the object to be stored (docs, images, videos, applications etc). Security is paramount in services and is taken care of here in two ways: Key authentication and SSL. The key authentication will work for most storage requirements.

Like many other storage services, S3 too provides a feature to expose a public URL for the object stored.

On the whole, the idea is to shield infrastructure details for companies to get an idea implemented quickly.

Business Model

Unlike, most business models of web 2.0, which are “free”, Amazon charges for data transfer and storage. But the charges seem nominal and the pricing details are well described. There is no option to try the service before signing up though. So, one needs to provide credit card details before using the service (I realized this when I got an exception that said so!).

When I first heard of the service, I thought the idea was a great innovation but was apprehensive about its implementation. Having toyed with the service for a while now, my apprehensions have disappeared. It is a great idea and an equally good implementation. Very valuable for startups.