Re: hello

[Jon Crowcroft <J.Crowcroft@cs.ucl.ac.uk>]

the problem it solves is the address _collision_ Problem - since all
SSM G,s are specific to their S's, you don';t have to run a
distributed algorithm or directory to do the allocation since there
are no clash type problems. i.e. the malloc architecture is
sidestepped.

In message <3AB1D1E5.A00CABCF@irisa.fr>, Ali Boudani typed:

 >>I was wondering that SSM (before was the SImple Muticast or EXPRESS) use
 >>a channel like (S, G) and the drafts are saying that this is the
 >>solution to the address allocation problem because each group is
 >>represented with the souce address also.  but why??
 >>figure this : 2 groups (having the same address) want to recieve from
 >>the same source. the (simple multicast or EXPRESS or SSM) will consider
 >>them as the same group
 >>So in this case any (S, G) = (S, G1) .... (S, Gn) and there is no sense
 >>in the adress group of the recievers.
 >>at the same time if G2 (the recievers in the G2 but having the G
 >>multicast address group) for example want to recieve from S2 , in that
 >>case all the G1.... Gn will recieve the data.
 >>
 >>am I right ??
 >>thanx
 >>

 cheers

   jon