Asymptotically Optimal Load Balancing Topologies

From MaRDI portal




Abstract: We consider a system of N servers inter-connected by some underlying graph topology GN. Tasks arrive at the various servers as independent Poisson processes of rate lambda. Each incoming task is irrevocably assigned to whichever server has the smallest number of tasks among the one where it appears and its neighbors in GN. Tasks have unit-mean exponential service times and leave the system upon service completion. The above model has been extensively investigated in the case GN is a clique. Since the servers are exchangeable in that case, the queue length process is quite tractable, and it has been proved that for any lambda<1, the fraction of servers with two or more tasks vanishes in the limit as Noinfty. For an arbitrary graph GN, the lack of exchangeability severely complicates the analysis, and the queue length process tends to be worse than for a clique. Accordingly, a graph GN is said to be N-optimal or sqrtN-optimal when the occupancy process on GN is equivalent to that on a clique on an N-scale or sqrtN-scale, respectively. We prove that if GN is an ErdH{o}s-R'enyi random graph with average degree d(N), then it is with high probability N-optimal and sqrtN-optimal if d(N)oinfty and d(N)/(sqrtNlog(N))oinfty as Noinfty, respectively. This demonstrates that optimality can be maintained at N-scale and sqrtN-scale while reducing the number of connections by nearly a factor N and sqrtN/log(N) compared to a clique, provided the topology is suitably random. It is further shown that if GN contains Theta(N) bounded-degree nodes, then it cannot be N-optimal. In addition, we establish that an arbitrary graph GN is N-optimal when its minimum degree is N−o(N), and may not be N-optimal even when its minimum degree is cN+o(N) for any 0<c<1/2.












This page was built for publication: Asymptotically Optimal Load Balancing Topologies

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6289169)