Integrated Performance Evaluation Approach based on an Advanced Traffic Modeling by Traffic Zones: A Case Study in Public Safety Network

摘要

The best traffic modeling of networks prevents an unexpected bottleneck traffic load, an appropriate allocation of network resources (proper number of link, channels, servers), and enhances performance, capacity of networks. Non-uniform traffic behaviors of networks that have random spikes, long tails make difficulties to detect and analyze traffic. In this paper, we present a Mixture of Lognormal Distribution Model (MLDM) based Quality of Service (QoS) evaluation technique that determines unobservable hidden traffic behaviors of networks using observable traffic data, and QoS parameters, their evaluation by traffic zones using incoming Key Parameter Identifications (KPIs). The posterior parameters of MLDM is estimated using the Bayesian probability method, and Lagrangian Multiplier techniques. The zoned traffic results illustrate that the developed approach can help to determine QoS parameters by traffic zones in networks. For the approach, we deal with two algorithms of the approach that can work together for the advanced approach using traffic data of Public Safety Network (PSN).