Latency-rate downlink packet scheduler para redes LTE

Abstract

The Long Term Evolution (LTE) standard is the currently leading technology used in 4G mobile networks around the world. Its basic architecture consists of three main components: a network core, base station, and user devices. User devices can request services available on the Internet, such as voice services, video and other not continuous data transmission services as Web pages. The access to these services are managed by the base station, which has the function of performing the resource scheduling, with a multiple access network technology and guarantee the quality of service to the user. Therefore, one of its main challenges is the support to services that require low latency and bandwidth. Among these, we highlight the video services for their constant growth related to the data consumption in mobile networks. Inside this context, this work presents the Latency-Rate Downlink Packet Scheduler (LR-DPS) for the traffic scheduling of downlink resources to guarantee the maximum delay requirements for variable rate traffic. Considering that the resource allocation in LTE networks is a complex optimization problem, the main contribution of the work is a new scheduling architecture with QoS for maximum delay. Three hierarchical steps introduce the proposal and formulation of the problem. In the first step, a token bucket bound the traffic. In the second step, the total time allocation (TTA) problem is solve, resulting in a total time for data transmission and the individual rates for guarantee the requested maximum delay. The third step allocates the data to resource blocks into a balanced way, meeting the requirements of the previous step. The performance evaluation carried out in the network simulator ns-3, with realistic scenarios that explore LTE network features and the LR-DPS model. In order to evaluate the performance, was employed the Round Robin and Channel and QoS Aware schedulers. The results of the simulations performed with different variable rate video (VBR) traffic presenting that the LR-DPS met the QoS requirements when the other schedulers exceeded the requested maximum delay by up to 90%.