[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

dc-nets



> Marc Ringuette, and Yanek Martinson are some of the folks on this list who
> are interested in working on DC-Nets....Yanek even claims to have a
> primitive one running on his local machine and is interested in volunteers
> to test it on a larger basis.

I did write a dc-net system in perl that uses e-mail as the transport
mechanism.  It worked quite well when all the "participants" were
on the same machine, and with immediat emessage delivery, no batching
or queuing.  As soon as I attempted to run it with a few people
over the net, there were severe synchronization problems.  E-mail
messages are not guaranteed to arrive in the same order as they
were sent, or to arrive to all the addressees within the same time
frame.  DC-nets work only if all the participants see the same set
of messages in the same sequence.

I have since decided that e-mail is not a very good transport
mechanism for a dc-net and am experimenting with tcp/socket
connections instead.  That seems to be working pretty well, but
has quite a high load on the network.

If someone does want to work on some sort of a synchronisation
protocol for the e-mail based dc-net, let me know and I will send
you the perl code.

The tcp version is not ready enough to give to anyone, and I have
been quite busy with a few other (only marginally cypherpunks-related)
projects and can't spend very much time on the dc-net system.