Locally Optimal Load Balancing

Laurent Feuilloley (1, 2) , Juho Hirvonen (2) , Jukka Suomela (2)

1 ENS Cachan - École normale supérieure - Cachan
2 Department of Information and Computer Science [Aalto]

This work studies distributed algorithms for locally optimal load-balancing: We are given a graph of maximum degree ∆, and each node has up to L units of load. The task is to distribute the load more evenly so that the loads of adjacent nodes differ by at most 1. If the graph is a path (∆ = 2), it is easy to solve the fractional version of the problem in O(L) communication rounds, independently of the number of nodes. We show that this is tight, and we show that it is possible to solve also the discrete version of the problem in O(L) rounds in paths. For the general case (∆ > 2), we show that fractional load balancing can be solved in poly(L,∆) rounds and discrete load balancing in f(L,∆) rounds for some function f, independently of the number of nodes.

Domains: Distributed, Parallel, and Cluster Computing [cs.DC]

Laurent Feuilloley, Juho Hirvonen, Jukka Suomela. Locally Optimal Load Balancing. DISC 2015, Toshimitsu Masuzawa; Koichi Wada, Oct 2015, Tokyo, Japan. ⟨10.1007/978-3-662-48653-5_36⟩. ⟨hal-01207214⟩