Packet buffering: randomization beats deterministic algorithms
STACS'05 Proceedings of the 22nd annual conference on Theoretical Aspects of Computer Science
Hi-index | 0.00 |
Packet losses in the current networks take place because of buffer shortage in a router. This paper studies how many buffers should be prepared in a router to eliminate packet losses in the context that an on-line scheduling algorithm in the router must decide the order of transmitting packets among m queues each of which corresponds to a single trafics tream. To exclude packet losses with a small amount of buffers, the maximum queue length must be kept low over the whole scheduling period. This new on-line problem is named the balanced scheduling problem (BSP). By competitive analysis, we evaluate the power of on-line algorithms regarding to the prevention of packet losses. The BSP accompanies tasks with negative costs. Solving an on-line problem which admits tasks with negative costs is our main theoretical contribution. We prove a simple greedy algorithm is ø(logm)-competitive and nearly optimal, while the ROUND ROBIN scheduling cannot break the trivial upper bound of m-competitiveness. Finally, this paper examines another balancing problem whose objective is to balance the delay among the m traffic streams.