Using Tailored Failure Suspectors to Support Distributed
F. Cosquer, L. Rodrigues, and P. Veríssimo
Selected sections of this report were published in the Proceedings of
the 7th IASTED/ISMM International Conference on Parallel and
Distributed Computing and Systems, Washington (DC), USA, October 1995.
This paper presents an approach to effectively support
cooperative applications using tailored failure suspectors. Using
a group communication subsystem, it is shown how failure
suspectors can be configured to model the requirements/semantics
of cooperative applications thus avoiding ad-hoc system decisions.
This approach is highly relevant in the context of large scale
distributed systems like the Internet, where communication high
variance and unpredictable delays increase the probability of
incorrect failure detection. Applications are presented
illustrating how failure suspectors are configured and possibly
combined with new feedback techniques in order to implement more
powerful cooperative environments.
Also available extended
report (gzip postscript), (pdf) .