Preemption policy in QoS-enabled networks: A customer centric approach

Preemption is a key technique to provide available and reliable services to high priority connections in a QoS-enabled network. Preemption technique is governed by a policy which makes the decision about which connections to preempt when resource scarcity is experienced. Existing preemption policies consider preemption rate, priority of connection and preempted bandwidth as the deciding criteria. In this paper, two additional criteria, namely, user satisfaction and revenue index are introduced in formulating an objective function that defines preemption policy as an optimization problem. We propose a model for the network enterprise to calculate an estimated level of user satisfaction of preemption enabled call connections in the context of service continuity and incorporate this information into preemption policy. Simulation results show that the proposed policies achieve improved customer perceived satisfaction and revenue index compared to the existing policies while preserving good system utilization, preemption rate and priority level of preempted calls. The improved user satisfaction and revenue index indicate higher prospects of revenue return and make the proposed policies highly attractive to the network providers as well as beneficial to the customer.