Today Amazon has announced its Computing on Demand offering through a service called Amazon Elastic Compute Cloud (Amazon EC2). The service is in limited beta. Basically, the service is about Amazon taking your hosting (or computing) needs on a on-demand basis. There pricing structure seems to include storage (S3), data transfer(Internet) and computing costs. The computing costs are calculated on hourly basis giving the flexibility to start/stop the Amazon Machine Image on a on-demand basis. It would be interesting to see how the hosting providers react to this especially shared hosting.
I would predict that the pre-configured AMI instances offerings will include LAMP (Linux, Apache, MySQL, PHP), LAMP with Tomcat (Java/J2EE), various flavors of Linux distributions, etc. similar to VMWare Virtual Appliances.
Since its a Computing solution my curiousity is to learn about what different system performance metrices/resource information will EC2 service be exposing. By this I mean, Will they be exposing real-time load on the servers? How about memory, processing power utilization related hardware information? It will also be interesting to see if there will be offerings for some form of analytics / reporting regarding the AMI instances. Or are these opportunities to be exploited by developers? Lets wait and watch.
It would be interesting to see the possibilities of services/products built based on EC2. Some of the services offered for S3 are here.
Updated: I liked this EC2 Review from Limited Beta participant, includes screenshots.