Amazon S3 Shakes up Web Storage

Unlike most “park your files here” storage services, Amazon S3 (short for Simple Storage Service) was developed to cater to a broad customer base, with special attention given to Web developers, fledgling ecommerce sites, and yes, even small businesses.

Essentially, S3 is as robust as Amazon.com, since the company’s Web Services division “takes Amazon’s software and infrastructure and basically repackages them for developers to use,” says Dave Barth, Product Manager S3, Amazon Web Services. Which is not to say that the company simply tied a bow on its internally developed Web services and called it a day. There’s the matter of added hardware, for starters.

S3 relies on many clustered storage servers spread across several geographically scattered sites. Though certain specifics are kept under wraps, the service relies on an infrastructure that has proven itself over the years as it is patterned after the guts that powers much of the ecommerce giant’s online operations.

Amazon has long conquered the challenges of providing reliable service and achieving massive scalability. An Amazon.com outage is a rare occurrence these days, even during peak buying times. And pricing is pegged at $0.15 per GB a month of stored data and $0.20 per GB transferred, which compares favorably to otherwise budgeting, building and maintaining a storage platform of one’s own, according to the company.

Of Objects, Keys and Buckets
Forget file structures, as you know them. “S3 is not a file system in the traditional sense of the word.” Indeed, this system may seem a foreign concept to people who keep their data in directories populated with folders, subfolders and files. Instead, S3 customers are given “a flat object space where every object is associated with it; there is no hierarchy.”

S3 uses keys that provide pointers to which customers can grant a naming convention to emulate folders. And buckets, as the name suggests, works as a domain space &#151 Amazon also uses it as a basis for billing.

An object can be of any size, provided that it doesn’t exceed 5 GB. Barth explains that this is a practical boundary, though there is “no limitation within the system under the interfaces.” Rather, it boils down to how much data can be reliably transferred over HTTP and SOAP online.

Barth notes that customers from a wide array of industries are adopting S3. In this regard, he feels that S3 has achieved its goal as a flexible online storage platform. With a little over a month since its debut, the Web developer community has been quick to take S3 for a spin and explore ideas on how to cost-effectively serve up data.

Interested parties can try out Amazon S3 here. For a look at the community that’s springing up around the service (and other Amazon Web Services offerings), visit the Amazon Web Services Blog.

Adapted from enterpriseitplanet.com.