Workload Alleviation Scheduling Framework to Alleviate Negative Performance Impact of Intermediate Data Skew in Small-Scale MapReduce Cloud

Tzu Chi Huang, Kuo Chih Chu, Jia Huei Lin, Guo Hao Huang, Ce-Kuen Shieh

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

A MapReduce cloud becomes the essential platform in the cloud computing infrastructure today. Because applications may process input data with different algorithms and logics to produce intermediate data, a MapReduce cloud may suffer intermediate data skew by unevenly distributing intermediate data among nodes at run time. When intermediate data skew happens, a MapReduce cloud not only idles nodes to waste computation resources but also prolongs the application execution progress to hurt user experiences in cloud computing. Instead of the existing solutions that assume many available idle nodes and use computation resources in a loose way, a MapReduce cloud can use the Workload Alleviation Scheduling Framework (W ASF) proposed in this paper to alleviate the negative performance impact of intermediate data skew in a small-scale MapReduce cloud by smartly utilizing computation resources. Besides, a MapReduce cloud is verified with popular applications in experiments to have the outstanding performance improvement with W ASF when intermediate data skew happens.

Original languageEnglish
Title of host publication2018 International Conference on System Science and Engineering, ICSSE 2018
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781538662854
DOIs
Publication statusPublished - 2018 Nov 1
Event2018 International Conference on System Science and Engineering, ICSSE 2018 - New Taipei City, Taiwan
Duration: 2018 Jun 282018 Jun 30

Publication series

Name2018 International Conference on System Science and Engineering, ICSSE 2018

Other

Other2018 International Conference on System Science and Engineering, ICSSE 2018
CountryTaiwan
CityNew Taipei City
Period18-06-2818-06-30

All Science Journal Classification (ASJC) codes

  • Control and Systems Engineering
  • Control and Optimization

Fingerprint Dive into the research topics of 'Workload Alleviation Scheduling Framework to Alleviate Negative Performance Impact of Intermediate Data Skew in Small-Scale MapReduce Cloud'. Together they form a unique fingerprint.

Cite this