Abstract: This unconventional talk consists of two independent topics: classical networks, and combinatorics.
(1) In classical networks, performance metrics such as round-trip delay and packet reordering are useful in diagnosing problems and attacks. However, these metrics are expensive to measure. And due to the volume of the traffic, detection should take place in the data plane. The first part of the talk will focus on efficient data-plane algorithms for measuring delay and reordering. In particular, we will see how leveraging probabilistic techniques helps in working with hardware restrictions on memory size and memory accesses.
(2) Enumerating polycubes is a fundamental combinatorial problem relevant to the study of percolation. In the second part of the talk, we will start with formulae for enumerating d-dimensional polycubes by volume and dimension, with a fixed perimeter defect. Moreover, we will see that for any fixed dimension d, the generating function that enumerates polycubes with a fixed defect is rational, and its denominator is a product of cyclotomic polynomials.
Bio: Yufei Zheng is a PhD student in the Department of Computer Science at Princeton University. She was a recipient of the Irwin and Joan Jacobs Fellowship. Yufei has worked on vastly different topics, with network algorithms and combinatorics being the two recurring themes. She is currently shifting her focus to quantum networks.