An Adaptive Failure Detection Protocol Export

@inproceedings{citeulike:3373624, abstract = {The detection of process failures is a crucial problem system designers have to cope with in order to buildfault-tolerant distributed platforms.Unfortunately, it is impossibleto distinguish with certainty a crashed process froma very slow process in a purely asynchronous distributedsystem.This prevents some problems to be solved in suchsystems.That is why failure detector oracles have been introduced to circumvent these impossibility results.This paper presents a relatively simple protocol that allowsa process to "monitor" another process, and consequentlyto detect its crash.This protocol enjoys the nice propertyto rely as much as possible on application messagesto do this monitoring.Differently from previous processcrash detection protocols, it uses control messages onlywhen no application messages is sent by the monitoring process to the observed process.This protocol has noteworthyfeatures.When the underlying system satisfies the partialsynchrony assumption, it actually implements an eventuallyperfect failure detector (i.e., a failure detector of the classusually denoted 3 P).Moreover, if the average observedtransmission delay is finite and the upper layer applicationterminates within a bounded number of steps for any failuredetector in 3 P after the failure detector becomes "perfect",then, when run with the proposed protocol, it alsoterminates correctly.These properties make the protocolattractive:it is inexpensive, implementable, and powerful.The paper also describes performance measurements of animplementation of the protocol.}, address = {Washington, DC, USA}, author = {Fetzer, Christol and Raynal, Michel and Tronel, Frederic}, booktitle = {PRDC '01: Proceedings of the 2001 Pacific Rim International Symposium on Dependable Computing}, citeulike-article-id = {3373624}, citeulike-linkout-0 = {http://portal.acm.org/citation.cfm?id=882475.883554}, citeulike-linkout-1 = {http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.15.362}, isbn = {0-7695-1414-6}, keywords = {2001, adaptive, adfs, failure\_detection, failure\_detectors, implementation, model}, posted-at = {2008-10-04 12:13:34}, priority = {2}, publisher = {IEEE Computer Society}, title = {An Adaptive Failure Detection Protocol}, url = {http://portal.acm.org/citation.cfm?id=882475.883554}, year = {2001} }

TY - CONF ID - citeulike:3373624 L3 - citeulike-article-id:3373624 N2 - The detection of process failures is a crucial problem system designers have to cope with in order to buildfault-tolerant distributed platforms.Unfortunately, it is impossibleto distinguish with certainty a crashed process froma very slow process in a purely asynchronous distributedsystem.This prevents some problems to be solved in suchsystems.That is why failure detector oracles have been introduced to circumvent these impossibility results.This paper presents a relatively simple protocol that allowsa process to "monitor" another process, and consequentlyto detect its crash.This protocol enjoys the nice propertyto rely as much as possible on application messagesto do this monitoring.Differently from previous processcrash detection protocols, it uses control messages onlywhen no application messages is sent by the monitoring process to the observed process.This protocol has noteworthyfeatures.When the underlying system satisfies the partialsynchrony assumption, it actually implements an eventuallyperfect failure detector (i.e., a failure detector of the classusually denoted 3 P).Moreover, if the average observedtransmission delay is finite and the upper layer applicationterminates within a bounded number of steps for any failuredetector in 3 P after the failure detector becomes "perfect",then, when run with the proposed protocol, it alsoterminates correctly.These properties make the protocolattractive:it is inexpensive, implementable, and powerful.The paper also describes performance measurements of animplementation of the protocol. SN - 0-7695-1414-6 AD - Washington, DC, USA TI - An Adaptive Failure Detection Protocol PB - IEEE Computer Society T2 - PRDC '01: Proceedings of the 2001 Pacific Rim International Symposium on Dependable Computing KW - 2001 KW - adaptive KW - adfs KW - failure_detection KW - failure_detectors KW - implementation KW - model AU - Fetzer, Christol AU - Raynal, Michel AU - Tronel, Frederic PY - 2001/// UR - http://portal.acm.org/citation.cfm?id=882475.883554 ER -