A Control Based Solution for Integrated Dynamic Capacity Assignment, Congestion Control and Scheduling in Wireless Networks

This paper deals with the resource management problem of guaranteeing a desired QoS (Quality of Service) to connections supported by wireless subnetworks, while efficiently exploiting the capacity of the relevant air interface. In the literature (as well as in real implementations), this very complex problem is "split" in several subproblems; this paper focuses on three strictly related resource management subproblems, namely dynamic capacity assignment, congestion control and traffic scheduling: dynamic capacity assignment refers to the partition among the terminals (which are the traffic sources) present in the considered wireless subnetwork of the available (valuable) air interface capacity, congestion control refers to the decision about the portion of the offered traffic (i.e. the traffic generated by the terminals) which can he admitted (i.e. stored) into the terminal queues (the admitted traffic will be eventually transmitted over the air interface) and the portion which has to be discarded to avoid congestion, scheduling refers to the decision about the traffic portion, stored in the terminal queues, which can be transmitted over the air interface (i.e. to the transmission priorities which are granted to the admitted The paper presents' an original, simple, and efficient control-based solution for jointly (i.e. in a fully integrated way) coping with the above-mentioned three resource management subproblems. Such solution is based on the proper distribution of simple control laws among some local controllers (located in the terminals generating the offered traffic) and a global controller (in charge of coordinating the local ones) and on the idea to use proportional feedback control in order to steer the overall system towards an ideal equilibrium at which desirable performance is achieved. In particular, the global controller is in charge of periodically updating this ideal equilibrium, which is a function of the traffic presently offered to the considered wireless subnetwork, whilst the local controllers are in charge to drive, as fast as possible, the overall system towards the current ideal equilibrium.