DMCA
Dominant resource fairness: Fair allocation of multiple resource types (2011)
Cached
Download Links
| Citations: | 146 - 15 self |
BibTeX
@TECHREPORT{Ghodsi11dominantresource,
author = {Ali Ghodsi and Matei Zaharia and Benjamin Hindman and Andy Konwinski and Scott Shenker and Ion Stoica},
title = {Dominant resource fairness: Fair allocation of multiple resource types},
institution = {},
year = {2011}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
We consider the problem of fair resource allocation in a system containing different resource types, where each user may have different demands for each resource. To address this problem, we propose Dominant Resource Fairness (DRF), a generalization of max-min fairness to multiple resource types. We show that DRF, unlike other possible policies, satisfies several highly desirable properties. First, DRF incentivizes users to share resources, by ensuring that no user is better off if resources are equally partitioned among them. Second, DRF is strategy-proof, as a user cannot increase her allocation by lying about her requirements. Third, DRF is envyfree, as no user would want to trade her allocation with that of another user. Finally, DRF allocations are Pareto efficient, as it is not possible to improve the allocation of a user without decreasing the allocation of another user. We have implemented DRF in the Mesos cluster resource manager, and show that it leads to better throughput and fairness than the slot-based fair sharing schemes in current cluster schedulers. 1
Keyphrases
dominant resource fairness multiple resource type fair allocation max-min fairness resource type fair resource allocation drf allocation possible policy mesos cluster resource manager user cannot desirable property different demand different resource type current cluster scheduler share resource
