Right. I have links to all the projects of this type that I could find
at http://pobox.com/~kragen/decentralized-networks.html, HYPHOS among
them. But the HYPHOS web pages don't have a heck of a lot of
information; no source code, no protocol specifications, etc. Just a
few pretty pictures. I wrote to one of the HYPHOS people; he averred
that the code (to simulate these guys) had been written, but didn't
respond to my request for further information. (Maybe he had better
things to do than respond to some lowbrow who wanted to play with his
toys.)
Has anyone else figured out how to solve the technical problems of
munchkins? The ones I see are:
1. routing -- how do you figure out how to get packets from one place
to another? This problem is subject to several very difficult
constraints:
the solution must be robust in the face of a few hostile nodes
(think terrorist munchkins);
link bandwidth must scale with bisection bandwidth so that in
this scenario:
a is connected to b, c, and d;
b is connected to a, c, and d;
c is connected to all nodes;
d is connected to all nodes;
e is connected to c, d, and f;
f is connected to c, d, and e;
all connections are 10Mbps;
a can talk to f at 10Mbps at the same time b talks to e at
10Mbps. (They can't both try to route all their traffic
through, say, c.)
it must be scalable up to hundreds of billions of nodes,
eliminating most current Internet routing systems (like
OSPF).
2. latency -- how can I get any kind of reasonable latency when my
packets are being forwarded through several hops each mile?
3. data integrity with untrustworthy forwarding stations.
-- <kragen@pobox.com> Kragen Sitaker <http://www.pobox.com/~kragen/> Computers are the tools of the devil. It is as simple as that. There is no monotheism strong enough that it cannot be shaken by Unix or any Microsoft product. The devil is real. He lives inside C programs. -- philg@mit.edu