Modelagem e simulação estocástica aplicada ao planejamento de capacidade de sistemas gateway de voz

Abstract

This work presents a systematic approach, based on stochastic modeling and discrete-event simulation, to plan the capacity of voice gateway systems. The proposal was evaluated through a case study, related to the analysis and capacity planning of a real voice gateway system from a Federal University. The modeled system processes, in average, more than 1000 phone calls daily. Given that voice gateways are critical to organizations, connecting to different providers, and requiring different resources to complete simultaneous calls, the use of analytical models are not adequate to plan and analyze their capacity. In addition to the flexibility to deal with high complexity modeling scenarios, the visualization of computer simulation was important during the engineers knowledge eliciting, mainly in the model validation step. The simulation model parameterization was conducted through the characterization of real call record data from the PABX and voice gateway analyzed. For this purpose, it was necessary the use of parametrical and non-parametrical statistical methods. Based on the proposed approach, it was possible to identify alternative setup for the study case, which allowed optimizing the use of the voice gateway s resources, especially its E1 trunk, increasing its availability in 14% with no addition resources. This represents an increase of 2926 new daily calls without any investment.