Degin and Implementation Priority Based Weighted Fair Queue of Based on MapReduce Cluster
Tian Qi · Computer Knowledge and Technology · 2011
This paper proposes a degin and implementation priority based Weighted Fair Queue of based on MapReduce clusters,according to drawbacks of Hadoop's MapReduce scheduler.It can clearly distinguish the level of customer service,while ensuring a degree of fairness and high priority was not subject to a small majority of users sharing system resources.It can achieve better data locality,improve the performance of the system and parallel access,and setting the appropriate for users and weight to ensure special tasks,according to the mathematical model.