On 07/26/2014 04:42 PM, Juliusz Chroboczek wrote:
Tossacoin, i.e. random routing, is of course a valid routing algorithm
in itself. Isn't that the original reason why we had a TTL/hop count? It
would probably work quite well in a small homenet.
I call it hot potato routing, and I define it as choosing a random
interface other than the one through which the packet was received.
I once brought some actual potatoes to my lecture, and we implemented hot
potato routing; it turned out being a good introduction to the subject of
packet drops. (I don't think we ever laughed so much during one of my
lectures.)
I wonder if anybody has ever considered routing from a natural selection
perspective. You throw random mutations (router chaos monkeys!) into the
routing tables every once in a while and see if the packet progeny dies off
or becomes more fit. Perhaps the mutation rate is the result of feedback
from the packet mortality rate, and backhoes are the equivalent of
asteroids,
cf punctuated evolution.
(maybe we can borrow some v6 address bits for the genetic material? :)
Mike
_______________________________________________
homenet mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/homenet
