Large-scale parallel server system with multi-component jobs

A broad class of parallel server systems is considered, for which we prove the steady-state asymptotic independence workloads, as number servers goes to infinity, while system load remains sub-critical. Arriving jobs consist multiple components. There are job classes, and each may be one two types, determines rule according components add workloads servers. The model enough include special cases some popular queueing models with redundancy, such cancel-on-start cancel-on-completion redundancy. Our analysis uses mean-field process representation corresponding limits. In essence, our approach relies almost exclusively on three fundamental properties model: (a) monotonicity, (b) work conservation (c) property that, average, “new arriving workload prefers go lower workloads.”