Amazon Web Services today announced the release of Cloudfront, a poor man's global content distribution network deeply integrated with the existing Simple Storage Service (S3) offering from the company. Cloudfront is able to take content stored in the S3 system and replicate it globally between 14 different datacenters and to send users accessing the content to the closest distribution point to reduce latency and increase performance.

The service has obvious benefits for existing S3 customers who are using the storage space to host website or other content intended for widescale distribution; for some, the performance benefits accrued simply from using the fast and powerful S3 system itself have been significant enough. Cloudfront will allow them to take the next baby step toward improving their global content distribution with a system that is dedicated to providing improved performance, not one where it is simply an ancillary benefit.

In addition to the benefits to web content providers, Cloudfront also has the potential to address the needs of organizations attempting to use cloud-based services for less conventional activities. One of the significant and unanswerable knocks on cloud-based services as a replacement for services traditionally hosted internally has been performance; the Internet is a fickle place and even the most robust solutions for non-web-based SaaS offerings have lagged in responsiveness and usability behind self-hosted alternatives. It's possible that Cloudfront could find a niche caching data and improving performance for any number of cloud-based services well beyond the obvious webhosting or software distribution applications.

While I think it's an excellent addition to the AWS family and probably exactly what many existing S3 users need to take their next steps toward global domination, I don't think anyone should mistake the service for a full-fledged worldwide content distribution network. Amazon CTO Werner Vogels lists the fourteen locations in which content will be cached on his blog; while they are located world wide, they hardly represent a challenge to the massive network that Akamai, perhaps the best known CDN, has stretched throughout seventy countries so far. Comments on Werner's blog reflect some disappointment in that respect. It's a mistake to evaluate this service on those terms, however; it's not supposed to kill Akamai, it's supposed to be an easy and efficient tool for mid-level providers to leverage their existing S3 data with. In that venue, I expect it will prove popular and successful.