Dipartimento di Ingegneria informatica, automatica e gestionale

In this paper the problem of guaranteeing Quality of Service via optimal buffer allocation is considered: namely, we compute the optimal buffer with respect to a real-world dataset accounting for the expected data traffic volume for a predefined set of business users belonging to a mobile network scenario. Two distinct approaches are followed to tackle this issue, in a static and dynamic fashion, respectively: the former relies on nonlinear programming, while the latter relies on model predictive control. The proposed formulation in terms of optimal static buffer allocation enables the minimization of the wasted amount of data traffic in order to prevent users from paying the assignment of extra resources in addition to the available traffic bundle. Instead, optimal dynamic buffer allocation pushes resource optimization forward by enabling personalization: indeed, tracking the prediction of data traffic consumption for each user allows to satisfy personalized Quality of Service guarantees. Numerical simulations show the effectiveness of the proposed approaches in terms of buffer saving and decrease in Quality of Service mismatch.