FaultTolerance.ppt

Fault Tolerance  /~course/cs501/2008 Hongfei Yan School of EECS, Peking University 5/5/2008 Failure is not an option. It comes bundled with your software.“ (--unknown) You know you have [a distributed system] when the crash of a computer youve never heard of stops you from getting any work done.“ (--Leslie Lamport) Some real-world datapoints Outline Basic concepts Process resilience Reliable client-server communication Reliable group communication Distributed commit Recovery Fault handling approaches Fault prevention: prevent the occurrence of a fault Fault tolerance: build a component in such a way that it can meet its specifications in the presence of faults (i.e., mask the presence of faults) Fault removal: reduce the presence, number, seriousness of faults Fault forecasting: estimate the present number, future incidence, and the consequences of faults Design Goal (with regard to fault tolerance): Design a (distributed) system that can recover from partial failures without affecting correctness or significantly impacting overall performance 分布式系统设计出发点 一个进程P可能依赖不同计算机上其他进程提供的服务，如果那些进程由于出现错误或故障而失去联系，则P无法正常运行。 计算机死机，或许网络断开，或许对方负载太重，暂时无法提供所需的服务 如果选取一组计算机联合执行同一个任务，当个别或少数计算机出现错误和故障时，大多数计算机仍然能够正常地完成任务 E.g., 分布式复制系统 Dependability Being fault tolerant is stringly related to what are called dependable systems. Some properties of dependability: Availability Readiness for usage Reliability Continuity of service delivery Safety Low probability of catastrophes Maintainability How easy can a failed system be repaired Terminology Failure: When a component is not living up to its specifications, a failure occurs Error: The part of a components state that can lead to a failure Fault: The cause of an error. Fault types generally (in terms of their properties): Transient: occur once and then disappear Intermittent: occur, then vanish, then reappear Permanent: continues to exist Failure Models (in terms of their specifications) Crash failures: A component