Distributed load balancing using alternating direction method of multipliers

Load balancing is a major issue in networking, where the traffic load is effectively distributed across multiple paths of a network graph. Practical networks have recently grown larger in size and complexity and demand, which necessitates the notion of decentralized load balancing. This paper provides an algorithm for balancing the traffic load in a general network setting. Unlike most of state-of-the-art algorithms in load balancing context, the proposed method is fully distributed. The effective distribution of loads is realized through solving a convex optimization problem where the p-norm of network load is minimized subject to network physical constraints. We employ the Alternating Direction Method of Multipliers (ADMM) to solve the p-norm problem in a distributed fashion, where each node only exchanges information with its neighboring nodes. Numerical results show that the proposed algorithm requires very few communication steps to balance the traffic load on bottleneck for random network graphs.