Guaranteeing the quality of service in cluster-based Web systems
Krzysztof Zatwarnicki, Leszek Borzemski · Frontiers in artificial intelligence and applications · 2012
The paper presents a new method for scheduling and distribution HTTP requests in the cluster-based Web service. Its application enables guaranteeing the quality of the Web service on the required level, opposed to other Web services offering unpredictable and unreliable services of best-effort type. The proposed MLF (Mostly Loaded First) method helps to keep the page response time within established boundaries, in this way that at a heavy workload, page response time both for small and complex pages, would not exceed the imposed time limit. The proposed method applies adaptive algorithms and neuro-fuzzy models which enable the estimation of page response time. In the paper, both the method and the testbed are described together with the results obtained in the simulation experiment.