资源描述:
《分布式并行计算环境下gml空间数据划分策略及算法的研究》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、AbstractGMLislargelyusedbecauseofitssimplicity,half-strutted,interoperability,openness,generalityandflexibilityetc.AstheGISproblemsbecomemoreandmorecomplexandthescalebecomelarger,thetraditionalGISspatialdatastorageandspatialanalyzealgorithmcannotmeettheneedofmassdatastorageandspatia
2、lanalyze.However,thisproblemcanperfectlysolvedbyacomputingmodelwhichiscalledthedistributedandparallelcomputing.Theperformanceofthedistributedandparallelcomputingismostlydependedonthestrategyofthedatapartitioningwhilethecurrentdatapartitioningalgorithmsdoesnottakethespatialrelationsh
3、ipstobeconcerned.Therefore,thispaperresearchesonsomespatialdatapartitioningalgorithmthatisappropriateforGML,concerningdatabalanceoneachnode,adjacencyofspatialobjects,areabalanceandspatialrelationships,andsomeinnovativearchivesarereachedasfollows:Firstofall,pointoutthedisadvantageoft
4、hespatialdataportioningbasedontheHilbertcurveandK-meansclusteringalgorithm.Theformerperformsnotwellonbalanceoftheareaforspatialdataoneachnodeandthelattermaygetabadresultforthesakeofabadinitialcentroid.Secondary,combiningtheHilbertcurveandK-meansclusteringalgorithm,proposeanewGMLdata
5、partitioningalgorithm,whichtakestheloadbalance,adjacency,areabalanceandspatialrelationshipsintoconsideration.Finally,basedonthealgorithmproposed,designedtheGMLdistributedstoragesystem,andfinishedthedatapartitioningmoduleofdistributedparallelGMLstoragesystembasedonHadoopplatform.Veri
6、fiedthedatabalanceoneachnode.ComparedtheConcurrencyAcceleratorRatioofthisalgorithmwiththosebasedonOracleSpatialorK-meansclusteringalgorithm.ComparedthecommensurateareaquerytimeofthisalgorithmwiththatbasedontheHilbertcurve.Theresultprovedthatthisdatapartitioningalgorithmhasagoodloadb
7、alanceandparallelqueryefficiency.KeyWords:GML;distributedcomputing;parallelcomputing;datapartitioning;HadoopII目录摘要........................................................................IAbstract...................................................................II目录.................
8、...................