Accurate Active Measurement in the Internet and its Applications

Since the early days of ARPANET, network performance measurement has played a key role in the continuous enhancement and evolution of the Internet. During this research project first a high accuracy measurement infrastructure ensuring highly accurate and reliable timestamping and test traffic generation was developed, and second, once the proper infrastructure was available, existing measurement methodologies were improved and new measurement methods were proposed and verified. As the result of this, the thesis contributes both to the Internet measurement infrastructure and to the are of measurement analysis. A highly accurate active probing infrastructure is provided without the need for GPS, based on an alternative software clock for PC’s running Unix. With respect to clock rate, it’s performance exceeds common GPS and NTP synchronized software clock accuracy. It is based on the TSC register counting CPU cycles and offers a resolution of around 1ns, a rate stability of 0.1PPM equal to that of the underlying hardware, and a processing overhead well under 1μs per timestamp. It is scalable and can be run in parallel with the usual clock. A highly accurate Real-Time Linux based probe stream sender solution is also introduced. A generic sample path based analysis of a multiple hop model of a network route carrying probe traffic and cross traffic is given. It provides a unifying context for active and passive end-to-end measurement techniques, and a framework to enhance existing and to generate new estimation methods. Novel aspects include focusing on delay variation and allowing general probe stream definitions. All network effects are identified and four canonical measurable signatures are defined and illustrated. The new delay variation based route model allows the principles of packet-pair to be formalised and extended. This enables the effect of probe size to be evaluated, downstream noise understood, peak detection recognised as superior to mode or minimum based filtering, and new estimation methods to be proposed and evaluated. Using insight from the governing equations and simulation, it is shown how real measurements made over typical routes can be interpreted. Unexpected additional probe size dependencies were found, inspiring an extension of the route model to include lower layer headers. It is shown how the enhanced model accounts very well for the observed dependencies, allowing more