Advanced search
Start date
Betweenand
(Reference retrieved automatically from Web of Science through information on FAPESP grant and its corresponding number as mentioned in the publication by the authors.)

Load Index Metrics for an Optimized Management of Web Services: A Systematic Evaluation

Full text
Author(s):
Souza, Paulo S. L. [1] ; Santana, Regina H. C. [1] ; Santana, Marcos J. [1] ; Zaluska, Ed [2] ; Faical, Bruno S. [1] ; Estrella, Julio C. [1]
Total Authors: 6
Affiliation:
[1] Univ Sao Paulo, Comp Syst Dept, Sao Paulo - Brazil
[2] Univ Southampton, Southampton, Hants - England
Total Affiliations: 2
Document type: Journal article
Source: PLoS One; v. 8, n. 7 JUL 16 2013.
Web of Science Citations: 1
Abstract

The lack of precision to predict service performance through load indices may lead to wrong decisions regarding the use of web services, compromising service performance and raising platform cost unnecessarily. This paper presents experimental studies to qualify the behaviour of load indices in the web service context. The experiments consider three services that generate controlled and significant server demands, four levels of workload for each service and six distinct execution scenarios. The evaluation considers three relevant perspectives: the capability for representing recent workloads, the capability for predicting near-future performance and finally stability. Eight different load indices were analysed, including the JMX Average Time index (proposed in this paper) specifically designed to address the limitations of the other indices. A systematic approach is applied to evaluate the different load indices, considering a multiple linear regression model based on the stepwise-AIC method. The results show that the load indices studied represent the workload to some extent; however, in contrast to expectations, most of them do not exhibit a coherent correlation with service performance and this can result in stability problems. The JMX Average Time index is an exception, showing a stable behaviour which is tightly-coupled to the service runtime for all executions. Load indices are used to predict the service runtime and therefore their inappropriate use can lead to decisions that will impact negatively on both service performance and execution cost. (AU)

FAPESP's process: 10/02839-0 - Distributed applications testing: challenges in message passing paradigm
Grantee:Paulo Sérgio Lopes de Souza
Support type: Scholarships abroad - Research
FAPESP's process: 08/57870-9 - Critical Embedded Systems Institute
Grantee:Jose Carlos Maldonado
Support type: Research Projects - Thematic Grants