VoIP packet scheduling for LTE advanced

The LTE-A (Long Term Evolution Advanced) technology constitutes a significant step in the evolution of 3G systems towards 4G. The performance targets set out for L TE Advanced make it an ideal solution to take care of the ever-increasing demand for wireless broadband services. L TE Advanced also comes along with new features and techniques that do not exist in LTE (Long Term Evolution), which is the 3G technology that precedes it. Such features include Carrier Aggregation, Coordinated Multipoint Reception and Transmission (COMP), relays and others. With the introduction of such features, Quality of Service (QoS) support for multiple data services such as voice and other multimedia applications is further enhanced. The Packet Scheduler (PS) which is a key entity located in the base-station/eNodeB plays a prime role as a key feature of LTE's Radio Resource Management in increasing the system's data rate and providing support for the diverse QoS requirements of mobile services. The Packet Scheduler should be able to distribute radio resources to mobile User Equipments (UEs) such that the LTE network is able to adhere to its performance requirements. Using dynamic system level simulations, this study performs an evaluation of a given set of scheduling algorithms. It seeks to determine whether these algorithms are capable of providing Quality of Service for VoIP traffic in a scenario where VoIP and web traffic compete for the same resources. In this study, two sets of algorithms were taken into consideration: QoS differentiated and non QoS differentiated packet scheduling schemes. It is found that those that take into consideration delay requirements of VoIP service presented the best overall performance. A performance evaluation is further carried out to determine the most suitable algorithm supporting VoIP traffic. It was observed that as regards VoIP service, Modified Largest Weighted Delay First Algorithm (MLWDF) outperforms other packet scheduling algorithms by offering a much higher system throughput, supporting more users and guaranteeing fairness at a satisfactory level. Based on this, an evolved ML WDF algorithm which is an enhancement to the existing ML WDF scheme is proposed in this study. The scheme proposed aims to minimise the packet loss of VoIP traffic that arises due to the delay of most packets at the user queues in the base station exceeding the delay threshold. The scheme does this by taking into account computation of parameters like packet arrival rate, buffer space capacity and packet size to reduce the packet loss for VoIP traffic. In the proposed scheme, priority is offered for VoIP traffic by performing resource reservation for VoIP packets iv on the Component Carriers (small bandwidths of up to 20 MHz) available for scheduling. In this study, an investigation is carried out on the important problem of downlink resource allocation in recently enhanced LTE Advanced systems where a newly added feature Carrier Aggregation provides more flexibility in Radio Resource Management