Bo Zhang, Filip Krikava, Romain Rouvoy, L. Seinturier
{"title":"Self-Configuration of the Number of Concurrently Running MapReduce Jobs in a Hadoop Cluster","authors":"Bo Zhang, Filip Krikava, Romain Rouvoy, L. Seinturier","doi":"10.1109/ICAC.2015.54","DOIUrl":null,"url":null,"abstract":"There is a trade-off between the number of concurrently running MapReduce jobs and their corresponding map and reduce tasks within a node in a Hadoop cluster. Leaving this trade-off statically configured to a single value can significantly reduce job response times leaving only sub optimal resource usage. To overcome this problem, we propose a feedback control loop based approach that dynamically adjusts the Hadoop resource manager configuration based on the current state of the cluster. The preliminary assessment based on workloads synthesized from real-world traces shows that the system performance can be improved by about 30% compared to default Hadoop setup.","PeriodicalId":6643,"journal":{"name":"2015 IEEE International Conference on Autonomic Computing","volume":"16 1","pages":"149-150"},"PeriodicalIF":0.0000,"publicationDate":"2015-07-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"11","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2015 IEEE International Conference on Autonomic Computing","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ICAC.2015.54","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 11
Abstract
There is a trade-off between the number of concurrently running MapReduce jobs and their corresponding map and reduce tasks within a node in a Hadoop cluster. Leaving this trade-off statically configured to a single value can significantly reduce job response times leaving only sub optimal resource usage. To overcome this problem, we propose a feedback control loop based approach that dynamically adjusts the Hadoop resource manager configuration based on the current state of the cluster. The preliminary assessment based on workloads synthesized from real-world traces shows that the system performance can be improved by about 30% compared to default Hadoop setup.
京公网安备 11010802042870号
京ICP备2023020795号-1
分享
求助内容：
应助结果提醒方式：
扫码关注我们
