Improved queue-size scaling for input-queued switches via graph factorization

From MaRDI portal
(Redirected from Publication:5005035)



Abstract: This paper studies the scaling of the expected total queue size in an nimesn input-queued switch, as a function of both the load ho and the system scale n. We provide a new class of scheduling policies under which the expected total queue size scales as Oleft(n(1−ho)−4/3logleft(maxfrac11−ho,night)ight), over all n and ho<1, when the arrival rates are uniform. This improves over the previously best-known scalings in two regimes: Oleft(n1.5(1−ho)−1logfrac11−hoight) when Omega(n−1.5)le1−holeO(n−1) and Oleft(fracnlogn(1−ho)2ight) when 1−hogeqOmega(n−1). A key ingredient in our method is a tight characterization of the largest k-factor of a random bipartite multigraph, which may be of independent interest.











This page was built for publication: Improved queue-size scaling for input-queued switches via graph factorization

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5005035)