Self-Adaptive Fuzzy Mechanism to Load Balancing on the Web Services Provider

• Anderson Talon Institution Toledo of Education
• Kelton Costa Institution Toledo of Education
• Edmundo Madeira Unicamp

Resumo

Nowadays an increase in web services is noticed. The monitoring is essential to guarantee the quality of services. Monitoring is the basis for verifying violations but can also be used to predict a possible violation. This paper proposes a mechanism for load balancing in the web services provider. The monitor uses fuzzy techniques to predict a non-functional failure and acts not to allow that it happens. This failure may occur due to overhead on the provider. Results show an improvement in the accomplishment of e-contracts and a decrease in response times average. Comparing the fuzzy scheduling with other schedules is observed an improvement of up to 83.02% in the e-contracts accomplishment and a decrease of up to 96.21% in response times average. Also, using the fuzzy system, the provider overload was better balanced varying by a maximum of 12.04%, while for the other schedules the variation reached 68.99%. Analyzing the influence of the quantity of services on the provider, with the increase of the load, the fuzzy mechanism achieved a performance up to 61.67% higher. The results show that the mechanism proposed in this paper is promising to assist in load balancing on the provider.

Palavras-chave: Time factors, Monitoring, Delays, Contracts, Computer architecture, Linguistics, Fuzzy systems, Load Balancing, Self-Adaptive Monitoring, Web Service Monitoring, Fuzzy Monitoring, Pro-Active Monitoring