A polling model is a queueing model consisting of several queues, which are cyclically visited by a server. The server visits the queues according to some discipline, like $1$-limited (serve at most one customer in a visit) or exhaustive (serve a queue until it has become empty). Polling models find many applications in computer-communications, and also in other areas like maintenance, production-inventory systems, and signallized traffic intersections.
The first part of the talk contains a global introduction to polling systems, and a review of some of their key properties. In the second part of the talk I'd like to describe some recent and ongoing work with Kamil Kosi\'nski and Offer Kella (on joint queue length and joint workload distributions) and with Ivo Adan, Urtzi Ayesta, Josine Bruin, Brian Fralix, Vidyadhar Kulkarni, Maaike Verloop, Adam Wierman and Erik Winands (on various scheduling and optimization problems in polling systems).