A preliminary framework using bpnn for performability modelling and analyses of a kerberos server

This paper presents a preliminary framework using back propagation neural networks for performability evaluation of Kerberos servers. In the modelling approach, keys under pseudo-secure conditions are dynamically renewed and also, temporary interruption to link/server access is involved. This interruption has implications on mortification of system's performance. Since this effects the network communication's cost, an analytical model is created for evaluation. Analytical model's results provide understanding on the values of authentication key renewal(restoration) times, times passed between renewals(restorations) and failures(breakdowns) of the server. Additionally, another distinctive way of critical thinking and problem solving is suggested, and an intelligent system using back propagation neural networks is applied to detect performability modelling.