文档下载 文档下载导航
设为首页 | 加入收藏
搜索 请输入内容:  
 导航当前位置: 文档下载 > 所有分类 > Increasing Availability by Sacrificing Data Integrity – a Problem Statement
侵权投诉

Increasing Availability by Sacrificing Data Integrity – a Problem Statement

Distributed software systems are the basis for innovative applications (e.g. pervasive computing, telecommunication services, and control engineering). The key for achieving scalable and maintainable distributed systems is dependability, because otherwise

Proceedings of the 38th Hawaii International Conference on System Sciences - 2005

IncreasingAvailabilitybySacri cingDataIntegrity

–aProblemStatement*

M.Jandl,A.Szep,R.SmeikalViennaUniversityofTechnologyInstituteofComputerTechnologyGusshausstrasse27-29/384,A1040Vienna,Austriasmeikal@acm.org

{jandl|szep}@ict.tuwien.ac.at

K.M.Goeschka

ViennaUniversityofTechnologyInstituteofInformationSystemsArgentinierstrasse8/184-1,A1040Vienna,AustriaKarl.Goeschka@tuwien.ac.at

Abstract

Distributedsoftwaresystemsarethebasisforin-novativeapplications(e.g.pervasivecomputing,tele-communicationservices,andcontrolengineering).Thekeyforachievingscalableandmaintainabledistributedsystemsisdependability,becauseotherwisethecom-plexityofdistributionwouldleavethesystemuncontrol-lable.Hence,ourapproachaimsataconceptforopti-mizingdependability.Wepresentaresearchproject*andcontributewithreasonsmotivatingtherelevanceofourobjectivebyshowingreal-lifeapplicationscenar-iosandbyanalyzingrelevantrelatedwork.Further,wecontributewithoursolutionapproachandthekeyidea:Similartootherapproachesweusereplicationasmeanstoprovidetransparentfault-toleranceandper-sistence,butweareespeciallyfocusedonincreasingavailabilitybyrelaxingdataintegrityusingamixtureofasynchronousandsynchronousreplicationtechniques.

fact,dependabilityisanaspectofeverypartofthesys-tem.Hence,weaimataconceptforoptimizingdepend-abilityindistributedcomponent-basedsoftwaresys-tems.Oneimportantwayofintroducingfault-toleranceistoaddfunctionalitytothesoftware,whichisabletorespondtosuchfaultysystemconditions[1],wherefail-uresinlowerlevels(i.e.hardwarefailures)canbeclas-si edassitefailures(aparticularsiteisnotworking)andnetworkfailures(networkpartitions,whereagroupofsitescannotbereachedbutisstilloperating).Inor-dertoenablereuseandtofreedeployeranddeveloperfromdealingwithsuchfault-tolerancefunctionality,ithastobeencapsulatedinwell-de nedandcoherentsys-temparts.However,improvingsinglecomponentavail-abilityisinsuf cient,butratherfault-toleranceneedstobeincorporatedintothesystemarchitectureitself.Largeandcomplexapplicationsusuallydemandbothtransparentdistributionandpersistence,andprovenplatformslike(a)Sun’sEnterpriseJavaBeans(EJB)andtheJava2Enterpriseservices,or(b)Mi-crosoft’sComponentObjectModel(COM+,.NET)withservicesor(c)theOMG’sCORBAarchitecturewithitsservices,exist.Asmentionedabove,distributedsystemsneedtobefault-toleranttobene tproperlyfromtheirstrengths—however,thelistedplatformsdonotre ectthisrequirementsuf ciently,asdescribedlater.Ontheotherhand,bypromotingtheideaofopensystems,wedonotaimatanewtypeofmiddleware,butratherattheintegrationofourapproachwithexistingcomponentinfrastructuresandCOTS(commercialofftheshelf)products.Furthermore,thisisneitheranewmethodforsoftwaretestingorveri cation,butratheraninnovativemethodtodealwithfailuresofnodesandlinksindistributedsystems.

1.ProblemStatementandResearchAim

Today,dependabilityismainlytakenintoconsidera-tionforhighlyspecializedsafetycriticalapplications.Unfortunately,thecomplexityofdistributedsystemsingeneralleavesthesesystemsuncontrollableandun-manageable,unlessavailability,reliability,andfault-tolerancearealreadybuiltintodesign,architecture,andcomponentinfrastructure.Thereisnosuchthinglikea“fault-tolerance”-modulethatcanbeaddedlateron.In

*SupportedbytheEuropeanCommunity(FP-6IST).Project“DependableDistributedSystems(DeDiSys)”,Contr.No.004152.

0-7695-2268-8/05/$20.00 (C) 2005 IEEE

1

第1页

你可能喜欢

热门文档

相关文档

站点地图 | 文档上传 | 侵权投诉 | 手机版
本站所有资源均来自互联网,本站只负责收集和整理,均不承担任何法律责任,如有侵权等其它行为请联系我们.
文档下载 Copyright 2013 doc.guandang.net All Rights Reserved.  -  email