Search papers, labs, and topics across Lattice.
This paper investigates the behavior of finite-time queue peaks in generalized switches within stochastic networks, focusing on how constrained service resources affect scheduling policies like MaxWeight. The authors reveal that while a square-root envelope characterizes peak behavior under certain conditions, this shifts to logarithmic growth beyond a geometry-dependent threshold, influenced by self-normalization mechanisms. Their findings, supported by simulations, highlight the critical interplay between queue dynamics and capacity geometry, offering new insights into peak performance limits in network systems.
Beyond a certain threshold, queue peaks grow logarithmically rather than quadratically, fundamentally altering our understanding of scheduling efficiency in stochastic networks.
We study finite-horizon queue peaks in generalized switches, a standard stochastic-network model in which many queues share constrained service resources. Arrivals may be dependent, time-varying, and adapted to the past; the standing load condition is uniform interior slack, meaning the conditional mean arrival vector stays in a fixed contraction of the capacity region. We show that this slack reshapes the finite-time peak law for drift-minimizing scheduling policies such as MaxWeight. The square-root envelope that is sharp without slack persists only up to a geometry-dependent threshold; beyond that threshold, the running maximum grows only logarithmically with the horizon, both with high probability and in expectation. The mechanism is self-normalization: in the current queue direction, the projected fluctuation scale is normalized by the stabilizing drift scale. This removes capacity geometry from the logarithmic coefficient, while geometry remains in the threshold. Matching lower bounds show that both the logarithmic term and a geometric threshold are unavoidable. When finite-time state-space collapse is available, the threshold can be sharpened using local bottleneck geometry. For generalized input-queued switches, we obtain finite-time peak bounds with tight logarithmic coefficients. Simulations illustrate the two-phase envelope, local geometric refinements, and variance-sensitive improvements predicted by the theory.
