there is research from all sorts of people in this area. It falls under the term "Ad-hoc networking". Google "Ad-hoc mobile network" and perhaps add in "routing" for more specificity.
You have a different cost metric for the communications between nodes than the mobile people do - and they have all sorts of other things like device power and such to worry about. But here's a nodal point : http://www.ee.surrey.ac.uk/Personal/G.Aggelou/MANET_PUBLICATIONS.html I seem to remember that efficient routing for cluster sizes of up to 1000 nodes was well proven (this was research from AT&T). Some sort of group effort in this direction would be interesting. I would help. ~~m On Wed, 2002-03-27 at 13:08, [EMAIL PROTECTED] wrote: > I've been thinking a little lately about network topologies > and peer-to-peer. What reading I've done seems to indicate that > most networks either have no organizational structure to them > at all or have some sort of dictated hierarchy. But it's > possible to have quite a lot of organization without anything like > a hierarchy. > > Here's a simple example of what I'm talking about: > > I've got a network of 1 million nodes, each of which has > an address 0-999999. For now, we won't ask how one goes about > attaining an address. The network is divided into "clusters" > of 100 and "superclusters" of 10,000. I'm assuming there aren't > any persistent connections. Imagine my nodes address is 123456. > Imagine I wanted to query all the nodes in the network. > I would directly query all the nodes in my cluster > (all those with addresses 1234xx), one node in each cluster in my > supercluster (for example, I'd query all nodes with addresses of the > form 12xx56) and one node in each supercluster other than my > own (for example, all nodes of the form xx3456). > So I'd query 300 nodes directly (297 for the persnickety), > the 100 nodes in the other cluster would each have to make 100 > second generation queries, and the 100 nodes I contact in the > other superclusters would each have to make 100 second generation > queries > leading to third generation queries. > > Of course in practice I would have to make a lot more queries, > because some of the nodes would be unavailable. But the point is, > assuming all the nodes which are running forward queries properly, > I should only have to actually talk to 300 nodes as described > above and, more importantly, none of the nodes should ever be > subjected to a dupliacte query from me. > > Note that there's no hierarchy here, all nodes are treated equally. > Node 129956 is sort of acting as a "gateway" for me to > cluster 1299xx, but only for me (and 10000 other people). > > Of course, this kind of structure could be made with any number of > "levels" of clustering. > > I don't recall ever having read of this type of structure before, > but it seems so obvious that I'm sure it's been discussed before. > So is there a name for it? Does anyone use it? has it been > shown to be utterly worthless? > > Thanks, > George