[webscaleizfun]

I've seen lots of potentially usable p2p CDN ideas, from storing data in each visitors browser & using them to serve other visitors said static content, to projects like IPFS where you can have a distributed website. The ultimate problems are static asset storage & distribution, DNS (how'd you make a fault tolerant, censorship resistant DNS solution that is updatable in nearly real time), and any database component that a site may need.

Most of these are solved issues[1], but it breaks down to site operators don't want to put in the extra effort or risk (what if a page has assets not load, or worse, bad assets are loaded), and the apathy of the average user prevents a web of trust or installation of any extra software from occurring unless it is drop dead simple and requires zero maintenance.

As soon as someone can bundle up these technologies in a nice package for both site owners and end users, they'll have a killer solution to improve site reliability & reduce operating expenses (by offloading CDN costs to the end users).

[1] - https://github.com/redecentralize/alternative-internet

[pigpigs]

IPFS, in particular, plan to have native browser support via a Javascript dependency. There was also PeerCDN (and bunch of similar clones), but they all don't tend to stick around for long or get widely used even though the setup is supposedly one line.

[shacharz]

The problem with PeerCDN and other WebRTC based solutions is that they can't solve the DNS DDOS problem. Since the WebRTC API needs a server for the handshake to start communicating between 2 peers directly.

[webscaleizfun]

Eh, but the DNS server isn't necessarily the same as the bootstrapping server for the WebRTC API, and in most cases it generally would not be.