Type a uncertainty in jitter measurements in communication networks

Reliable quality of service measurements are crucial issues to monitor and evaluate the performance of computer and communication networks. To these aims, depending on the applications, several figures of merit have to be measured. They mainly concern with: throughput, available bandwidth, packet jitter, one way delay, round trip delay, and packet loss. To estimate these quantities two general classes of measurement systems (also known as “protocol analyzers”) can be considered: Special Purpose Architecture based Protocol Analyzer (SPAPA) and General Purpose Architecture based Protocol Analyzer (GPAPA). Both measurement solutions can be thought as composed by a reference data traffic generator and one or more measurement nodes. All these components are constituted by hardware and software sections which cooperate to achieve the measurement results. Focusing the attention on the uncertainty involved in the QoS indexes measurement, no information is provided by the manufactures for both SPAPA and GPAPA. In addition, the analysis of the literature in the field of computer networks testing, has highlighted that this important aspect is not adequately dealt with or that methodological studies are not yet available. This paper proposes a suitable approach to evaluate the type A measurement uncertainty of SPAPA and GPAPA systems. A general model to study the effect of the involved quantities of influence is presented and a method to quantify their effects is provided. The experimental results show the suitability and the effectiveness of the proposal.