E.L. Cecílio, A.P.M. Dumont, R.B. Correia, L.F.R.C. Carmo, L. Pirmez.
Performance management is currently done, typically, in a reactive fashion. This means that a QoS failure is notified only after its occurrence. This paper proposes a pro-active architecture for performance management. Variations of QoS parameters are monitored in real time and Tendency Alarms are generated when, possibly, failures are prone to occur. Then, pre-emptive actions are automatically executed in attempt to avoid the probable failures or to minimize their effects, if they happen to occur. A distributed management infra-structure based on active technology and Java is used in a real application scenario. Tests results are presented, thus permiting the evaluation of the efficacy and performance of monitoring and reaction time of the pertinent systems.
http://www.lbd.dcc.ufmg.br:8080/colecoes/sbrc/2003/013.pdf
Caso o link acima esteja inválido, faça uma busca pelo texto completo na Web: Buscar na Web