In queueing theory, a discipline within the mathematical theory of probability, a layered queueing network (or rendezvous network[1]) is a queueing network model where the service time for each job at each service node is given by the response time of a queueing network (and those service times in turn may also be determined by further nested networks). Resources can be nested and queues form along the nodes of the nesting structure.[2][3] The nesting structure thus defines "layers" within the queueing model.[2]
Layered queueing has applications in a wide range of distributed systems which involve different master/slave, replicated services and client-server components, allowing each local node to be represented by a specific queue, then orchestrating the evaluation of these queues.[2]
For large population of jobs, a fluid limit has been shown in PEPA to be a give good approximation of performance measures.[4]