Multi-timescale Fairness for Heterogeneous Broadband Traffic in Access-Aggregation Networks

We propose a throughput value function (TVF) based solution for providing multi time-scale (MTS) fairness for broadband traffic in access-aggregation networks. The primary goal of MTS fairness is a dynamic control of resource sharing that considers the usage history of the broadband connection. We present a flow level description of the multi time-scale throughput value function (MTS-TVF) based resource sharing. We provide dimensioning guidelines in traffic aggregation scenarios and present its simulation-based performance analysis. In the performance analysis, our focus is on overloaded systems involving both low load users (with temporally active traffic) and high load users (with heavy traffic, e.g. continuous multiple downloads). We find that the Quality of Experience (QoE) of low load users significantly increases when using MTS-TVF and it becomes similar to that of a lightly loaded system, while the change in the QoE is minimal for high load users.