Average case analysis of greedy routing algorithms on arrays

76Citations
Citations of this article
6Readers
Mendeley users who have this article in their library.

Abstract

In this paper, we analyze the average case behavior of greedy routing algorithms on arrays under a variety of assumptions. Overall, we find that certain greedy algorithms perform surprisingly well on average, For example, given an N × N array or torus where every node starts with one packet headed for a random destination, we show that some (but not all) greedy store-and-forward algorithms route every packet to its destination with only O(log N) delay per packet and maximum queuesize 4 with probability near 1. Moreover, the expected delay per packet is only a small constant, independent of N. We also extend the analysis to a steady state model of routing in which packets enter the network at random times. Provided that the overall arrival rate of packets to the network is less than 100% of the network capacity, we show that any packet encounters at most O(log N) delay with high probability. In addition, we show that the maximum size of a queue over a time span of T steps is O(log T/log N) with high probability. The results can also be extended to analyze the average case behavior of cut-through (or, flit-serial) routing under lighter loading.

Cite

CITATION STYLE

APA

Leighton, T. (1990). Average case analysis of greedy routing algorithms on arrays. In Algorithms and Architectures (pp. 2–10). Publ by ACM. https://doi.org/10.1145/97444.97448

Register to see more suggestions

Mendeley helps you to discover research relevant for your work.

Already have an account?

Save time finding and organizing research with Mendeley

Sign up for free