DATE: Thursday, July 10, 2008
TIME: 2:00pm – 3:00pm
Intel Research Pittsburgh, 4720 Forbes Avenue, CIC Building 4th Floor, Suite 410
NOTE SPECIAL TIME AND LOCATION
Dr. Darryl Veitch
University of Melbourne
Probing Convex Networks
Packet delay and loss are two fundamental measures of performance in the Internet. Using active probing to measure delay and loss has often led to methods which send probes, or probe patterns, in a 'Poisson' manner. This has been justified by invoking the PASTA property (Poisson Arrivals See Time Averages), which ensures that measurements made by Poisson observers are "unbiased."
Recent work, however, has questioned the utility of PASTA for probing and shown that, for delay measurements,
i) a wide variety of processes other than Poisson can be used to probe with zero bias and
ii) Poisson probing does not necessarily minimize the variance of delay estimates.
I will discuss optimal probing processes that, in the non-intrusive case, minimize the mean-square error of measurement estimates for both delay and loss. The optimality result is general, and only assumes that the target process we seek to optimally measure via probing, such as a loss or delay process, has a convex auto-covariance function. We use empirical datasets to demonstrate the applicability of our results in practice. Together, these results lead to explicit guidelines on designing the best probe streams for both delay and loss estimation. This is particularly important for loss, whose rarity means that it is important to make the best possible use of scarce data.
I was born and raised in Melbourne, Australia some time ago. I graduated with a Bachelor of Science Honours degree in 1985 from the (then) Applied Mathematics Department at Monash University. I obtained my Phd in 1991 from the Department of Applied Mathematics and Theoretical Physics of the University of Cambridge (England) in Dynamical Systems, under the supervision of Paul Glendinning. I then entered telecommunications by joining the Telecom Research Laboratories in Melbourne where I spent 3 years and began my interest in `fractal' traffic. In 1994 I left to spend a year at the research labs of France Telecom, the (then) CNET in Paris, working on long range dependent queueing in the group of Jim Roberts and Alain Simonian. After a brief stay of 3 months at the department of Teleinformatics of the KTH in Stockholm as a guest lecturer, I returned to France as an ingénieur expert, working in the internet group at INRIA Sophia Antipolis (near Nice) with Jean Bolot. At this time I began working on the applications of wavelet-based parameter estimation with Patrice Abry from the Signal Analysis group of the Ecole Normale Supérieure de Lyon. I was then a consultant to Bellcore (now Telcordia), New Jersey USA, for a period of 5 months, working in the group of Ashok Erramilli. I returned to Australia in January 1997 to join SERC, an Ericsson funded software and networking research group, as a Senior Research Fellow. The networking group moved to Melbourne University to form EMULab in January 2000, which I managed before its merger with CUBIN in 2002. In 2003-4 I was on sabbatical at Sprint ATL, working in the IP group.
Coordinator: Shellee Lank, firstname.lastname@example.org
or visit http://www.pdl.cmu.edu/SDI/