Packet buffering: randomization beats deterministic algorithms

  • Authors:
  • Markus Schmidt

  • Affiliations:
  • Institut für Informatik, Albert-Ludwigs-Universität Freiburg, Freiburg, Germany

  • Venue:
  • STACS'05 Proceedings of the 22nd annual conference on Theoretical Aspects of Computer Science
  • Year:
  • 2005

Quantified Score

Hi-index 0.00

Visualization

Abstract

We consider the problem of buffering unit value data packets in multi-queue network switches where each of the switch's input ports is equipped with a buffer of limited capacity. At these ports, packets arrive online and can be stored within the space limitations or must be discarded. Our objective is the maximization of the number of forwarded packets where, per time step, at most one packet from the set of buffers can be transmitted to the output port. In this paper, we give a technique for transforming any randomized algorithm for unit buffers into a randomized algorithm for buffers with arbitrary capacities while maintaining the competitiveness. We present the first randomized online algorithm that beats the deterministic lower bound of e/(e – 1) ≈ 1.58. It is 3/2-competitive and thus nearly matches the randomized lower bound of 1.46. For buffers with 2 queues having large capacities, we show a lower bound of 16/13 ≈ 1.23 for any online algorithm and prove that the competitive ratio of greedy algorithms is 9/7 ≈ 1.29, improving the best previously known upper bound of 3/2.