Performance Evaluation of MU-MIMO Under the Impact of Open Loop Traffic Dynamics

Abstract

Multi-user MIMO enhances throughput by simultaneously transmitting/receiving parallel data streams to/from a group of users. However, the throughput analysis does not ac-count for variable data traffic. The research objective of this project is to analyze the effect of variable traffic as it will change the system behavior. The two characteristic features of variable traffic viz. packet size variation and traffic burstiness have been considered for the analysis in this project. Their effects have been studied individually to provide an insight into the problem. Through simulations, we show that in certain scenarios the system performance of MU-MIMO deteriorates significantly due to both packet size variations and traffic burstiness individually even under ideal conditions with respect to channel variations, channel correlation, mobility, etc. Furthermore, we show that under bursty traffic with higher offered load the aggregate throughput first re-mains steady until a certain peak to average rate ratio and then deteriorates linearly instead of exponentially. Also, un-der high amount of traffic burstiness, the throughputs are in-dependent of the aggregation rates. A thorough analysis is provided to explain these two phenomena. This is the first research work that considers the impact of variable traffic on the performance of MU-MIMO. Therefore, the implications for MAC protocol design are significant.