P2P sharing is a fantastic way for users to share files, and cached content is merely made up of files and HTTP meta data. So why not create a peer-to-peer network that caches content for frequently accessed Web sites in the area. For example, users have a client running on their system that stores information about frequently accessed sites, and that information is aggregated in a local index over a P2P server in a region.

The regional index server has pointers to files on local P2P nodes, which are needed to render commonly accessed pages such as Google.com, Facebook, Twitter, and others that pulls the files locally from CDN clients locally in the area, rather than from the origin or even the CDN of the content provider. This way, a large amount of data can be distributed across many in a local geographic location, but instead of being served by a CDN, which one organization must pay for, the cost of performance is croudsourced.