On Thursday, Amazon Web Services introduced Elastic IP addresses and multiple Availability Zones, two new features for developers using Amazon's Elastic Compute Cloud (EC2) that make the AWS platform even more useful for tying dynamic computing infrastructure to Web sites.
"We give you an IP address which is mapped to your account," said Adam Selipsky, VP of product management and developer relations for Amazon Web Services. "You can then easily map the address to any EC2 instance."
An EC2 instance might be a virtual Web server for example, perhaps deployed in conjunction with Amazon Simple Storage Service (Amazon S3), Amazon SimpleDB and Amazon Simple Queue Service (Amazon SQS) for storage, computing, and query processing.
An Elastic IP address can be easily remapped to point to different server instances, as might be desirable when testing a new build of a Web application. Thus, Amazon is enabling a significantly more fault tolerant way to manage servers that would be available to developers running their own physical servers in a data center.
Amazon Web Services is also offering Availability Zones, which allow for EC2 instances to be run in multiple locations.
Using Availability Zones, the same Web application, for example, could be made available in the U.S. and in Europe and a failure in the EC2 instance serving one location would not affect the other instance. This sort of fail-over system no doubt would have been helpful in February when Amazon's S3 went down for several hours.
Despite such hiccups, Amazon Web Services appears to be growing. While Selipsky declined to provide comprehensive figures about the service, he did say that S3 had grown to 14 billion objects stored in January, up from 10 billion stored in October 2007. Amazon also recently said that for the first time, the bandwidth required for EC2 and S3 for the fourth quarter of 2007 exceeded the bandwidth required for all of Amazon's Web sites worldwide during that period.