A Tale of Tails: Asymptotics and Non-asymptotics in Load Balancing

Explore a technical lecture on queueing systems and load balancing presented by Prof. Siva Theja Maguluri from Georgia Tech. Dive deep into the complexities of studying queueing systems, examining various asymptotic regimes including many server, heavy traffic, and large deviations approaches. Learn about recent developments in prelimit results and rate of convergence characterization, with a specific focus on join-the-shortest-queue load balancing as a practical example. Understand how nonasymptotic tail bounds on queue lengths serve as a bridge between different asymptotic regimes, all demonstrated through the transform method. Benefit from the expertise of Prof. Maguluri, the Fouts Family Early Career Professor whose work in Control, Optimization, Algorithms, and Applied Probability has earned multiple accolades including the NSF CAREER award and Best Publication in Applied Probability award.