Optimizing Hadoop for the cluster.pdf

Optimizing Hadoop for the cluster.pdf

ID:33924141

大小:125.12 KB

页数:4页

时间:2019-02-28

Optimizing Hadoop for the cluster.pdf_第1页
Optimizing Hadoop for the cluster.pdf_第2页
Optimizing Hadoop for the cluster.pdf_第3页
Optimizing Hadoop for the cluster.pdf_第4页
资源描述:

《Optimizing Hadoop for the cluster.pdf》由会员上传分享,免费在线阅读,更多相关内容在学术论文-天天文库

1、OptimizingHadoopfortheclusterChristerA.HansenInstituteforComputerScienceUniversityofTromsø,Norwayhansen.christer@gmail.comABSTRACTbrie ysomerelatedwork.Section7concludesthispaperandsuggestssomefuturework.ThetotalnumberofclustersrunningHadoopincreasesev-eryday.Thereasonforthisisthatcomp

2、anieshavefoundasimplemodelthatworkswell.Themodelisbuilttoworkon2.TERMINOLOGYthousandsofmachinesandmassivedatasets.MapReduceTheHadoopopen-sourceframeworkconsistsofadistributedisascalableandfault-tolerantmodelthathidesallthedirty lesystemandtheprogrammingmodelitself.The lesys-workforthepr

3、ogrammers.SinceHadoopisbeinginstalledtemiscalledHadoopDistributedFileSystem(HDFS)andonmoreandmoreclusters,con gurationbecomesanimpor-issimilartotheGoogleFileSystem(GFS)[4].Hadoopistanttopic.Thisfactinspiredusto ndoutifweareablebasedontheMapReduceprogrammingmodeldesignedbytooptimizeHadoo

4、ptorunfaster.WehaverunexperimentsGoogle[3].Inthissectionweprovidesomehighleveltermi-ondi erentcon gurationsandfoundthatHadoopshouldbenologyregardingMapReduceandHadoop.con guredinawaythatitutilizestheresourcestheclusterAtypicalHadoopjobisasfollows:TheHadoopframe-hastoo er.Thedefaultcon g

5、urationofHadoophasprovenworkiscalledwithinaregularprogram.Theuserprogramtobeslow.Weuseourowncon gurationandCloudera'sthemapandreducefunctions,andsuggestshowmanymap-Con guratortoprovethatperformancemayincreasesignif-perandreducertasksthatshouldbeexecuted.Theinputicantlybyjustchangingaset

6、ofparameters. leissplitupinto64MBblocksbydefault(butiscon g-urable),andisdistributedamongtheslaves.The lesareKeywordsstoredonthesamemachinesthatexecutethemappertasks.Theresultofamaptaskisasetofintermediatekey/valueCon guringHadoop,Clouderapairs,thisdataisnotstoredinthedistributed lesyst

7、embutonthelocal lesystemoftheslave.Thelocationsof1.INTRODUCTIONtheseintermediatekey/valuepairsarepassedbacktotheCon guringHadooptoperformwellonaclusterislikemasterwhichisresponsibleforforwardingtheselocationstuningacar.Thereareseveralswitchesandknobsthattothereduceworkers.Whent

当前文档最多预览五页,下载文档查看全文

此文档下载收益归作者所有

当前文档最多预览五页,下载文档查看全文
温馨提示:
1. 部分包含数学公式或PPT动画的文件,查看预览时可能会显示错乱或异常,文件下载后无此问题,请放心下载。
2. 本文档由用户上传,版权归属用户,天天文库负责整理代发布。如果您对本文档版权有争议请及时联系客服。
3. 下载前请仔细阅读文档内容,确认文档内容符合您的需求后进行下载,若出现内容与标题不符可向本站投诉处理。
4. 下载文档时可能由于网络波动等原因无法下载或下载错误,付费完成后未能成功下载的用户请联系客服处理。