S3+EC2+X=The future of the Web
Just a quick braindump after researching viable options for bootstrapping a web startup: S3, EC2 and SQS really signals the dawn of a new era. This kind of technology enables people who know the social web to do what they do best, and focus less on scaling, storage, distribution, and all the other boring issues. The X that’s missing in the equation is a database in the sky (and persistent storage for EC2 instances, but that’s apparently in the works…). Mysql+Amazon in a partnership is my bet.
S3 prices were lowered yesterday, and as other competitors enter the field of cloud computing and infinite web-accessible storage we’ll see prices go down rapidly. I can’t wait to see a player mirroring the AWS API:s–cutting prices in half while they’re at it. As the market for pay-as-you-go web infrastructure matures, low-switching-costs-multisourcing will become possible and this in turn will drive down prices fast. Tight coupling between services–such as the free transit between S3 and EC2–will be key to survival for future infrastructure providers.
The next stage after EC2-style virtualization is of course to provide not only many instances of CPU:s but rather one monster of a machine where cores, RAM and disk scale transparently on demand. But that is a super hard computer science problem that some smart Stanford hackers are probably really busy with as I speak. One of the hard parts of building web apps today is really managing the transitions from a single box to a multi-tier architecture to database segmentation to etc, etc. Imagine if you could take your weekend-hack of a Rails app and put it out there with the possibility to scale it up to millions of users without having to think about these things. Abstraction, abstraction is the way to solve all hard problems in computer science. We are getting there. >>END Braindump