[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

Re: [tor-talk] Tor and P2P

> While we don't need a very complex p2p design (in short, we are mostly
> just talking about simple HTTP servers running on each device, behind
> a hidden service .onion), I am concerned in the long run about
> scalability and reliability of this.

Is, or can Tor be, useful or optimal? On which platforms? With what
P2P apps or traffic patterns?

Assuming these are the only cores available at this time...


Running a P2P app will require running one of the above cores. Which
means all peers already have access to each core's internal addressing
scheme, respectively: i2p, private IPv6, onion. And thus they have access
to each other.

So though Tor is unique in having exit facilities by design, those
facilities are not needed for a global P2P mesh to work.

Tor's uniqueness and its perhaps as yet uncharacterized ability to
handle certain P2P loads should lead to a comparative review of the
utility of other cores.

If you can speak IPv6, you can speak Phantom, which makes app
development incredibly simple. Written in C. Probably the lightest
weight core so far. Needs some review. Presents a native, node
specific, source address.

I2P has the most network experience so far with lots of P2P connections
going on, ala bittorrent. More work is needed to interface apps
with it. Written in Java. Does not present a native, node specific,
source address?

If you need to use exit for other things, you may not have room to
run multiple cores, provide some resources back, etc. So Tor would
make sense in that case. Does not present a native, node specific,
source address.

There are certainly more comparisons to be made.

> OnionCat can be seen as implementing P2P-over-HS. I never read a
> similarily strong statement about it before.

It does provide an IPv6 interface/layer, that is definitely a win since
you can use all the usual network programming methods and apps.

It is not a P2P app, only a shim for Tor and I2P. And it needs
further work to provide IPv6 to HS/I2P PKI authentication. I might
be inclined to sign all the traffic ala IPSEC before attempting
some ping pong scheme.

On a big scale, it might make sense for all the anonymous networks
to use interfaces and IPv6 node addresses for interop purposes.
Though there are some current problems there... Tor's 80 bits may
be too small, I2P's 256 is wider than can be packed under an IPv6/48.
tor-talk mailing list