How bad are wrong suspicions? towards adaptive distributed protocols

Sampaio, Luciano Menezes Bezerra; Brasileiro, Francisco; Cirne, Walfredo; Figueiredo, Jorge C. A. de

doi:10.1109/dsn.2003.1209965

Cited by 12 publications

(31 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Therefore, the issue of tuning these timeouts is a crucial point in the design of an efficient failure detector. Too small timeouts increase wrong suspicions, which, may harm the performance of the higher level consensus protocol [17]. On the other hand, too large timeouts increase the detection latency of the service, imposing heavy overheads to those runs on which failures occur [4,23].…”

Section: Related Workmentioning

confidence: 99%

“…In fact, in the simulations conducted, the failure detector module was replaced by a mock implementation that always returned an empty set of suspected processes, obviating the need for exchanging failure detection messages. Nevertheless, the simulation results presented in [17] show that even this very naive implementation of a slowness oracle is able to increase the performance of consensus in as much as 17% for a lightly loaded system.…”

Section: Raul Ceretta Nunesmentioning

confidence: 99%

“…This is a common scenario for many applications, including the emerging grid middleware infra-structure (see for example the scheduler service of the PARADIS system [10]). In this scenario, if the coordinator of the first round is correct but runs slow (because either its processor or its communication channels -or both -are overloaded), the performance of the protocol may suffer [17]. Sampaio et al [17] have proposed the use of slowness oracles to tackle this problem.…”

Section: Raul Ceretta Nunesmentioning

confidence: 99%

“…In this scenario, if the coordinator of the first round is correct but runs slow (because either its processor or its communication channels -or both -are overloaded), the performance of the protocol may suffer [17]. Sampaio et al [17] have proposed the use of slowness oracles to tackle this problem.…”

Section: Raul Ceretta Nunesmentioning

confidence: 99%

“…In the very simple implementation of a slowness oracle presented in [17], processes are classified as slow or fast based on the round trip delays of the messages they exchange during the past executions of consensus. There is a fixed threshold that is used to identify slow/fast processes.…”

Section: Raul Ceretta Nunesmentioning

confidence: 99%

See 4 more Smart Citations

Efficient and robust adaptive consensus services based on oracles

Sampaio¹,

Brasileiro²,

Nunes³

et al. 2005

J. Braz. Comp. Soc.

View full text Add to dashboard Cite

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Raul Ceretta Nunesmentioning

confidence: 99%

Section: Raul Ceretta Nunesmentioning

confidence: 99%

Section: Raul Ceretta Nunesmentioning

confidence: 99%

Section: Raul Ceretta Nunesmentioning

confidence: 99%

See 3 more Smart Citations

Efficient and robust adaptive consensus services based on oracles

Sampaio¹,

Brasileiro²,

Nunes³

et al. 2005

J. Braz. Comp. Soc.

View full text Add to dashboard Cite

show abstract

Adaptive Indulgent Consensus

Sampaio

Brasileiro

2005 International Conference on Dependable Systems and Networks (DSN'05)

View full text Add to dashboard Cite

Due to their fundamental role in the design of faulttolerant distributed systems, consensus protocols have been widely studied. In particular, design and performance issues of indulgent consensus is a research topic that has gained considerable attention. Most of these protocols are asymmetric in the sense that different participants can assume different roles during the execution of the protocol. Usually, there is a process that assumes a "special" role and the others cooperate with it to finish the computation. However, the asymmetric structure of indulgent consensus protocols has a performance pitfall, specially when processes and communication channels are subject to considerable variability in load. The problem is that such protocols use an a priori agreed process ordering to select the process to perform the "special" role. We advocate that adaptive indulgent consensus protocols can be constructed by the introduction of an adaptive process ordering module. In this sense, it is proposed a generic implementation for this module. Based on this generic module we provide implementations of both 3S-and Ω-based adaptive indulgent consensus protocols. Further, we investigate their performance by means of simulation and real experiments over a widely distributed system. The experimental results obtained show that the adaptive consensus protocols can outperform their non-adaptive counterparts in as much as 50%.

show abstract