Hi,

i am a beginner in networking area. during the study of adhoc networks, i have a question. if i have a 2 level hierarchical network, with N nodes where N is a large no. and x regions or groups, then what should be the best value for x. Should it be large or small. Also, if the nodes are large in number, is 2 level hierarchy beneficial?



Any help is appreciated

Yes you need the hierarchy, or layers as cisco calls them. With large numbers of nodes you need an access layer and then separate or collapsed core and distribution layers. I try to keep subnets/vlans to a /24 size or smaller but thats just empirical rather than theory backed

hi,
i mean that if i have an adhoc network which has 2 level hierarchy, which means that there would be N nodes divided into x groups or regions. the complexity in this case would be x.O(N/x). If N is a very large number, then keeping x a constant would not improve the asymptotic notation, so i think that x should be a function of N, for example x=N/constant. Is this the best we can do, or we can have x that can further improve the asymptotic notation? that is why i asked the best value of x.
Another question i have is that if we consider the hierarchy 2-level as a function of x and N is it beneficial or having a greater hierarchy (3,4,5-level) is better?