- Scalable to big meshes
This is tricky. The mesh routing schemes I have seen, and the discussions at the IETF (manet) etc. Seem to all have a hard cieling in the thousands of nodes. The only mesh routing protocol that seems to have a chance of scaling up to internet size is geographic routing. It's controversial. There is a privacy concern that can be solved with address encryption.
- Use automating routing on Internet to link distant meshes
This comes for free so long as the hosts inside the mesh, are given routable internet addresses. That's easiest if there is only a single internet gateway. If there's multiple internet gateways, then it's still possible, it becomes a multihoming problem. I think it's still possible to have a distributed scheme to make that work though.
Mario hi,
- check Manet type of routing protocols (AODV, DSR)
Done, I have also found newer protocols (GPSR, HSR, etc.)
GPSR (greedy perimeter stateless routing) relies on geographic addressing though. Is that an assumption you make? If so please allow address encryption, otherwise it's a bit privacy problem, that the IP address gives away exact coordinates.
- consider interplay between routing layer and underlying wireless technologies (802.11b, Bluetooth)
Done, it is better to abstract from level 2.
Agreed! Forget about layer 2 !!! ;-)
simon
-- www.simonwoodside.com -- 99% Devil, 1% Angel
-- general wireless list, a bawug thing <http://www.bawug.org/> [un]subscribe: http://lists.bawug.org/mailman/listinfo/wireless
