Authors:
Anderson Francisco Talon
1
and
Edmundo Roberto Mauro Madeira
2
Affiliations:
1
University of Campinas (UNICAMP) and Faculty FGP (FGP), Brazil
;
2
University of Campinas (UNICAMP), Brazil
Keyword(s):
Fuzzy Monitoring, Self-Adaptive Monitoring, pro-Active Monitoring, Web-Service Monitoring, e-Contract Violation.
Related
Ontology
Subjects/Areas/Topics:
Advanced Applications of Fuzzy Logic
;
Artificial Intelligence and Decision Support Systems
;
Enterprise Information Systems
;
Scheduling and Planning
Abstract:
The rise of web services have become increasingly more visible. Monitoring these services ensures Quality
of Service and it is the basis for verifying and potentially predicting e-contract violations. This paper
proposes a fuzzy scheduling mechanism that attempts to predict a possible e-contract violation based on
historical data of the provider’s services. Consequently, there is a self-configuration on the architecture that
changes service priority, making the provider processes the high priority services before low priority
services. This prediction can also helps the self-optimization of the architecture. A decrease of e-contract
violations can be observed. Though it is not always possible to predict a failure, the architecture is capable
of self-healing by using recovery actions. Comparing the fuzzy scheduling with others known in the
literature, an improvement of 31.52% in the e-contracts accomplishment is observed, and a decrease of
35.59% in average response time was
achieved. Furthermore, by using the fuzzy scheduling, the overload of
the provider was better balanced, varying at most 8.43%, while the variation in other scheduling
mechanisms reached 41.15%. The results show that the fuzzy scheduling mechanism is promising.
(More)