There are errors in my comments on multiple clients checking the same mailbox. It's easy to compute min(A,B) where A and B are independent identically exponentially distributed random variables. Just another exponential! :) It's harder if they've minimum and maximum delays though.
In any case, Christian Grothoff found an easy solution when I mentioned
the problem :
Fix the maximum number n of machines allowed on the mailing list.
Select a waiting time random variable T. If there are k participants
then choose {x_1,...,x_k} such that x_1 + .. + x_k = n. Tell the ith
machine to check as if it were x_i independent machines, either by
working out \min_j=1^{x_i} T_j or by simply using x_i independent
timers. If k changes, then reassign the x_i.
Jeff
signature.asc
Description: This is a digitally signed message part
_______________________________________________ Messaging mailing list [email protected] https://moderncrypto.org/mailman/listinfo/messaging
