Re: [freehaven-dev] RFC: Design for an anonymous network

[Roger Dingledine <arma@mit.edu>]

On Mon, Mar 05, 2001 at 01:22:16PM -0500, Michael J Freedman wrote:
> A wishes to transmit some MSG to M.
> A builds the packet as such:
> 
> ROUTE_CREATE_PACKET
> 
> Routing Information Header
> 
> [K_s_1, K_d_1 {K_s_2, K_d_2 (K_M, K_d_2, K_d_1, K_A)_PK_M}_PK_2]_PK_1
 
This appears to not give 2 any means of communicating back to A. That
means that if A chooses a path A -> 1 -> 2 -> M, and M is down, then
2 won't have any recourse but to close the connection to 1 and think
to itself "sorry, bud." At which point 1 says to itself "uh, gosh"
and closes the connection to A, at which point A knows *something*
screwed up and has no idea what.

I think this is an ok failure mode. We should just be aware of it.



Also, this 'routing information header' is the point at which A sends
more parameters to the various hops along the path. A can send such
things as "please plan to use 20KB/s traffic for me". This will allow
each hop along the way to choose whether to accept the request. It's
got an easy hook for including payment down the road.

It also makes it kind of clumsy from an economics point of view, since
until the connection starts getting used, it's nearly free, and there's
no telling how long it will wait idle until somebody 'plugs' into it.
Thoughts?
 
--Roger