Ideas
A few things to get the hamster wheel turning
Resilient Web Applications
As of Spring 2009, Amazon has been offering load-balancing and auto-scaling for its EC2 instances. With these highly desirable features as well as elastic IP addressing, persistent storage, and excellent network latency, EC2 has situated itself as the ideal platform for developing and deploying large scale web apps. With load balancing, all of your instances will be uniformly utilized, and with the auto-scaling, your cluster of instances can grow and shrink based on necessity. These features not only make an application highly resistant to failure, but also optimize resource consumption to keep your service charges at a minimum.
Ad-Hoc Data Center
Since Amazon can offer computation on an hour-by-hour basis, the Elastic Computing Cloud is perfect for on-demand computing needs. If you have a one-shot data processing job, it is very cost effective to configure an ad-hoc cluster which can be created with thousands of processing nodes in a matter of minutes. There has also been research into using EC2 as a platform for running MPI (Message Passing Interface). With EC2 instances having up to 8 64-bit cores each, the potential for computation in the cloud is nothing to overlook.
Content Delivery Network
Improve load times by serving your static content (javascript, images, CSS, et al.) off of S3 with Amazon CloudFront. CloudFront enables your static content to be highly available and very fast by distributing a cache of your content across several network locations on the “edge” of Amazon’s cloud. The speed improvement brought by a CDN is two-fold: first, it offloads the burden of serving a large number of static files from Apache, and second the network latency for content served through a CDN is much lower than that of traditional web servers. Furthermore, with CloudFront you get a personalized distribution URL that can be used to configure your server so your URLs still point to your domain (think cdn.yourdomain.com instead of d213ka94fsl1sd3.cloudfront.amazon.com).
Large Data Sets
In addition to many public data sets hosted on Amazon, users can upload large data sets on to Amazon S3 or onto Amazon EBS volumes. S3 is capable of storing individual files up to 5GB with no maximum for the number of files; EBS volumes can be sized from 1GB to 1TB.
