1数字系统的故障诊断与可靠设计

[e-mail];://test.ict.ac.cn“”“”1.2.Fault-TolerantComputerFault-TolerantComputingDependableComputing()1946-1957IBM650,UNIVAC,Whirlwind,SAPO/TMRCPU:SAPO(1956)EDVAC1949SAGESAPO1957-19641964-1970SSI/MSIV1964⌧JPLSTAR1969⌧BellESSNo.11965⌧1970~LSI/VLSIESSNo.3a1975TANDEM161975AMDAHL470V/61976FTSC1978IBM43001979FTMP1980SIFT19801980~///InternetVLSI/SoC1980-2000~dependability.org/wg10.4/IFIP10.4WG()(DSN)A.Avizienis,etal,“Basicconceptsandtaxonomyofdependableandsecurecomputing,”IEEETrans.onDependableandSecureComputing,Jan.-March2004,Vol.1,No.1,pp.11-33••••FMEAFMECAFMEAMIL-STD-1543MIL-STD-785SJ2166-82FMEA/FMECAFMEA/FMECAI/O1.…2.…3.…4.…5.…6.…………112232132431425415265162761728718298193092031102132112233IIIIIIabcde6.()7.()Basicterminology(I)failure:acomponent'sinabilitytoperformconformitsspecification–(notethatspecificationscanbewrong!)error:causeofafailure,i.e.,awronglogicalsystemstatefault:anomalous(physical)componentcondition:–internal:designormanufacturingfault,damage,aging–external:harshconditions(water,lightning,temperature),radiation,electromagnetism,misuseExplanationandexample?Basicterminology(II)error=therehasbeenafaulttherehasbeenafault≠errorafault:–canbe“sensitized=error–butcanalsoremainlatent(foralongtime)anerror:–canbe“propagated”=failure–butcanalsoremaininvisible1.Permanent2.Intermittent3.Transient1.;2./3.4.5.6.7.8.9.10.“”“”“”“”DualActiveHotStandbyPrimaryServerStandyServerA.Avizienis,etal,“Basicconceptsandtaxonomyofdependableandsecurecomputing,”IEEETrans.onDependableandSecureComputing,Jan.-March2004,Vol.1,No.1,pp.11-3320083107:00pmS106:(),,2000