Sampling and Inference in Network Measurement
Friday October 18, 2002
Host: Bud Mishra, email@example.com, 212-998-34645
Communications network providers collect diverse operational data from their networks for the purposes of planning, engineering, and control. However, finding the right view of network behavior from this data is challenging: phenomena of interest are not always directly observable; their presence may be revealed only through joining different data sets; and the enormous volumes of data present obstacles for management and the extraction of detail.
This talk will describe (i) tomographic inference of network usage and performance not directly observable; (ii) importance sampling to manage the analysis of large network data sets; and if time permits (iii) some applications to usage-based charging.
Nick Duffield holds the B.A. in Natural Sciences (Physics and Theoretical Physics, 1982) and the Certificate of Advanced Study in Mathematics (1983) from the University of Cambridge, U.K. He was awarded the Ph.D. (1987) by London University, U.K., for a study of dynamics and stability of non-equilibrium phase-transitions in biological systems. He held postdoctoral positions in Heidelberg, Germany, and Dublin, Ireland, before joining the faculty of the School of Mathematical Sciences in Dublin City University, Dublin, Ireland in 1991. He moved to AT&T Labs-Research in 1995, where he is currently Technology Leader in the Network Management and Performance Department. His research focuses on performance measurement, inference and analysis of communications networks. He is active in the IETF and was charter Chair of its working group on Packet Sampling.