Uncertainties Analysis in RTT Network Measurements: the GUM and RFC Approaches

Measurement of the Internet performances is a complicated process because of its large diffusion, heterogeneity, and decentralized nature. To date, request for comments (RFC) of the Internet Engineering Task Force (IETF) represents the de facto official protocol standard. Since each network metric, or set of metrics, can be evaluated by means of several measurement procedures, the optimal measurement methodology is the one that minimizes the measurement uncertainty. To this purpose, RFC expresses the need of supplying with documents and of quantifying the contribution of the measurement uncertainty. However, indications for quantitatively representing measurement uncertainty are generically addressed. A general purpose procedure to be followed for characterizing the quality of a measurement result is given in the ISO guide "Guide to the expression of uncertainty in measurement". In this paper a comparison between approaches proposed by the RFC and by the GUM for the evaluation of the measurement uncertainty contribution of the round-trip time (RTT) network parameter is addressed. The experimental evaluation of the RTT by means of three different methodologies is also presented and each corresponding measurement uncertainty is analyzed