Distributed Offline Load Balancing in MapReduce Networks
2013 (English)In: 2013 IEEE 52ND ANNUAL CONFERENCE ON DECISION AND CONTROL (CDC), IEEE conference proceedings, 2013, 835-840 p.Conference paper (Refereed)
In this paper we address the problem of balancing the processing load of MapReduce tasks running on heterogeneous clusters, i.e., clusters with different capacities and update cycles. We present a fully decentralized algorithm, based on ratio consensus, where each mapper decides the amount of workload data to handle for a single user job using only job specific local information, i.e., information that can be collected from directly connected neighboring mappers, regarding their current workload and capacity. In contrast to other algorithms in the literature, the proposed algorithm can be deployed in heterogeneous networks and can operate asynchronously in both directed and undirected communication topologies. The performance of the proposed algorithm is demonstrated via simulation experiments on large-scale strongly connected topologies.
Place, publisher, year, edition, pages
IEEE conference proceedings, 2013. 835-840 p.
Emerging control applications, Agents and autonomous systems, Linear systems
Control Engineering Computer Systems Communication Systems
IdentifiersURN: urn:nbn:se:kth:diva-136169ISI: 000352223501010OAI: oai:DiVA.org:kth-136169DiVA: diva2:675501
52nd IEEE Conference on Decision and Control; Florence, Italy, 10-13 December, 2013
QC 201512082013-12-042013-12-042015-12-08Bibliographically approved