[p2p-hackers] Introducing Pagekite - a pragmatic attempt at a p2p web

Michael Blizek michi1 at michaelblizek.twilightparadox.com
Tue Jan 18 15:33:58 EST 2011


On 16:35 Tue 18 Jan     , Bjarni Rúnar Einarsson wrote:
> Hello people!
> I thought some here might be interested in a project I've been working on
> for a few months, named Pagekite.
> (tl;dr: https://pagekite.net/downloads/)
> Pagekite is a very pragmatic attempt to enable more p2p-like behavior on the
> WWW by making it really easy for people to run publicly visible HTTP (or
> HTTPS) servers from personal and/or mobile devices. The system is not true
> "p2p" in that it relies on intermediate servers (front-ends) to do
> application layer routing (proxying/tunneling). but the hope is that it will
> bring at least some p2p features to the WWW. The connection between a web
> server and a front-end is an outgoing (from the web-server) TLS tunnel,
> which means it goes right through most firewalls and NAT, and can even be
> tunneled over SOCKS or Tor. The front-end is similar to a traditional HTTP
> load-balancer which does vhost-name based back-end selection, except the
> configuration is completely dynamic and all the traffic is sent over said
> encrypted tunnels.

This sounds very interesting. I guess that you can do something similar with
some of the VPN services out there, but creating a native protocol for HTTP
might be way more efficient. It would be possible to run lots of web sites
with a single frontend and a single IP. It should also be possible to reduce
the traffic between the frontend and the backend by doing some caching and

What is particularly interesting is that running web sites can easily be done
anonymously. However, this will put some strains on the front end operaters.
But then the connection between the end user and the front end should really
be encrypted.

> The primary inspiration for this project was actually the exact same
> reasoning as I noticed on the recent thread about p2p social networking,
> about how proximity to the end-user and the end-user's hardware and
> peripherals should be leveraged to add features to free/distributed social
> networks that the cloud-based incumbents would have a hard time matching. I
> think Diaspora and Appleseed and all the other social web projects are
> horribly handicapped by their decision to compete on Facebook's home turf,
> if you will. Facebook have shown themselves to be very technically competent
> and with their resources and user-base, they'll be incredibly hard to beat
> at their own game. Simultaneously, the true p2p networks are handicapped by
> the requirement that people install specialized clients to talk to their
> friends which leads to a chicken-and-egg problem for adoption and growth.

Embedded linux based NAS systems have IMHO still a lot of potential. I would
really like to see them transform into "home servers" which can take back
control from the "cloud" to the end users. I think that your piece could be
quite important.

> Thoughts?

Very nice...

programing a layer 3+4 network protocol for mesh networks
see http://michaelblizek.twilightparadox.com

More information about the p2p-hackers mailing list