Asymptotically optimal load balancing in large-scale heterogeneous systems with multiple dispatchers

Abstract We consider the load balancing problem in large-scale heterogeneous systems with multiple dispatchers. introduce a general framework called Local-Estimation-Driven (LED). Under this framework, each dispatcher keeps local (possibly outdated) estimates of queue lengths for all servers, and dispatching decision is made purely based on these estimates. The are updated via infrequent communications between dispatchers servers. derive sufficient conditions LED policies to achieve throughput optimality delay heavy-traffic, respectively. These directly imply many previous local-memory heavy traffic. Moreover, results enable us design new optimal Finally, heavy-traffic also sheds light recent open question how schemes using delayed information.