Analytic modeling of load balancing policies for tasks with heavy-tailed distributions

We present an analytic technique for modeling load balancing policies on a cluster of servers conditioned on the fact that the service times of arriving tasks are drawn from heavy tail distributions. We propose a new modeling methodology for the exact solution of an M/Hk/1 server and illustrate its use for modeling two distinct load balancing policies in a distributed multi-server system. Our analytic results provide exact information regarding the distribution of task sizes that compose the waiting queue on each server and suggest an easy and inexpensive way to provide load balancing based on the sizes of the incoming tasks.

• Type: Conference
• Year: 2000
• Where: WOSP
• Authors: Alma Riska, Evgenia Smirni, Gianfranco Ciardo
• Keywords: Balancing, Heavy Tail Distributions, Load Balancing Policies, Software Engineering, WOSP 2000