Self-Stabilizing Load Balancing Algorithms

A distributed system consists of a set of loosely connected state machines which d o not share a global memory. The global state of the system depends on the states of each ofthe processes in the system. The set of global states can be split up into two categories, legal and illegal. A desired property of the system is that regardless of the initial state of the system, the system automatically converges to a legal state in a finite number of steps. Also, if an error occurs in the system causing the system to be put into an illegal state, the system will again correct itself and converge t o a legal state an a finite amount of time. If the system is able t o d o this, it is called a self-stabilizing system. A load balancing algorithm attempts t o distribute the tasks around the system so that no processors are extremely overloaded or underloaded. This paper presents two self-stabilizing load balancing algorithms. The system is an arbitrary network of processors, and once the system stabilizes, the load will be balanced around the system.