I think that is called a load leveler. It needs to keep track of the status of the resources, and the capacity of each one, when means that it must be able to detect when a processor goes down. It should be able to monitor the queue length of each processor and distribute the requests to the least busy processor. If no processors have a queue of requests waiting, requests could just be assigned in rotating order. Finally, the system needs to be able to recover when one processor goes down, which means that it must store the contents of each request, and if the response times out, it must send the request to another processor. Alternatively, fault recovery can be achieved by redundancy, sending every request to two processors and using the first response. This should improve worst case response time, but obviously it requires twice as many processors.
Don On Apr 15, 5:09 am, "w.s miller" <[email protected]> wrote: > suppose we have n processing units.we have an incoming traffic which is to > be > distributed among them.but we have to take in mind that load is to be > equally distibuted and if one of the processing units goes down ,then we > have to distribute the traffic to rest of the processing units.Also when > the processing unit which is down comes up,then the traffic distribution is > to be handled .... > > please suggest any suitable data structure or algorithm.... > > thanks -- You received this message because you are subscribed to the Google Groups "Algorithm Geeks" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/groups/opt_out.
