The Power of Two Choices in Randomized Load Balancing
IEEE Transactions on Parallel and Distributed Systems
Heavy traffic resource pooling in parallel-server systems
Queueing Systems: Theory and Applications
Geometric generalizations of the power of two choices
Proceedings of the sixteenth annual ACM symposium on Parallelism in algorithms and architectures
Optimal Routing In Output-Queued Flexible Server Systems
Probability in the Engineering and Informational Sciences
Asymptotic Optimality of Balanced Routing
Operations Research
Heavy traffic optimal resource allocation algorithms for cloud computing clusters
Proceedings of the 24th International Teletraffic Congress
| Hi-index | 0.00 |
This paper considers the problem of routing Poisson arrivals to N parallel servers under the condition that the system is heavily loaded. We propose a scheme in which a proportion of arrivals are routed randomly, while the others are routed to one of two neighbouring queues using load information. We show that this scheme, which exploits a limited amount of load information and takes into account locality considerations, achieves performance close to that of a routing policy which requires complete load information. In addition, we show that this scheme has a diffusion scaled queue length process that is the same as if all of the servers were pooled with a single queue (in other words, no routing decision need be made). Our insights provide an additional option in load balancing, complementing the related work on the power of two choices.