Amazon may not be as prolific with their spinoff projects as Google but what they lack in number they make up for in sophistication. One of my favorites was the The Mechanical Turk (Beta) (sometimes it seems like the entire computer industry is stuck in beta) an ambitious and definitely novel service for human augmented grid computing. The recently announced Amazon EC2 (Amazon Elastic Compute Cloud) takes the company into a new realm of grid-style service.
EC2 is an inexpensive, highly scalable, pay-as-you-go, grid solution that uses Amazon’s own server infrastructure. Each cloud “instance” “predictably provides the equivalent of a system with a 1.7Ghz Xeon CPU, 1.75GB of RAM, 160GB of local disk, and 250Mb/s of network bandwidth” with root access to each one allowing you to take complete control of each one. To use EC2 you have to create an Amazon Machine Image (AMI) that contains the entire applications infrastructure you need to run or you can use one of Amazon’s pre-configured AMIs. The AMI and its data are stored in Amazon’s Simple Storage Service (Amazon S3). The costs of using EC2 are indeed reasonable at “$0.10 per instance-hour consumed (or part of an hour consumed), $0.20 per GB of data transferred outside of Amazon (i.e., Internet traffic), and $0.15 per GB-Month of Amazon S3 storage used for your images (charged by Amazon S3).” The potential of this service is enormous. Just consider, processing a 100GB dataset for 8 hours using 100 instances would cost around $15! This makes EC2 the cheapest render farm for small computer animation and graphics companies. It will be interesting to see where Amazon takes this service and what uses the market makes of it.
