When trying to decide where to host our soon to be released products, Amazon’s EC2 cluster was one of the first services we looked at. It was in beta mode, but we liked the flexibility it provided. Unfortunately, at the time, there were several problems that kept us from using the service. Two of the big ones were no static IPs and lack of a static hard drive. Both of these problems have recently been dealt with, which is fair as they are still in beta development. Amazon has even rolled out an impressive customer service plan in response to Google’s app engine. In the end, there was still one problem which amazon could not overcome.

Amazon’s default virtual machines contained only a rudimentary setup. We had to figure out on our own how to deal with load balancing, app server configuration, database clustering, etc. As a start-up company trying to develop our first app, we had limited experience in these fields. The Amazon EC2 service seems to be designed specifically to help small to medium sized companies expand, so these are issues that many of their potential users probably have to face.

A feature that could have been the deciding factor for us would have been a larger range of pre-configured instances. Perhaps a collection of template networks to address commonly used network configurations. For example, a virtual instance with the load balancing apps already installed and initially configured. This way, we’d only have to modify a single configuration file to add load balancing to our network. Another useful example would be a database instance. Once configured, such a virtual machine could have several instances deployed to create a database cluster. Many start-up companies don’t have the resources to design their own solutions and configurations to these common problems. Really, they shouldn’t have to. Amazon could really expand their client base by having these template networks for common configurations to allow rapid development of new apps.