English

Derandomized Load Balancing using Random Walks on Expander Graphs

Probability 2024-12-20 v2 Networking and Internet Architecture Performance Systems and Control Systems and Control

Abstract

In a computing center with a huge amount of machines, when a job arrives, a dispatcher need to decide which machine to route this job to based on limited information. A classical method, called the power-of-dd choices algorithm is to pick dd servers independently at random and dispatch the job to the least loaded server among the dd servers. In this paper, we analyze a low-randomness variant of this dispatching scheme, where dd queues are sampled through dd independent non-backtracking random walks on a kk-regular graph GG. Under certain assumptions of the graph GG we show that under this scheme, the dynamics of the queuing system converges to the same deterministic ordinary differential equation (ODE) for the power-of-dd choices scheme. We also show that the system is stable under the proposed scheme, and the stationary distribution of the system converges to the fixed point of the ODE.

Keywords

Cite

@article{arxiv.1901.09094,
  title  = {Derandomized Load Balancing using Random Walks on Expander Graphs},
  author = {Dengwang Tang and Vijay G. Subramanian},
  journal= {arXiv preprint arXiv:1901.09094},
  year   = {2024}
}

Comments

20 pages