Analysis of channel quality reporting mechanisms and their impact into 4G and beyond technologies

Resumen

Modern broadband mobile networks such as 4G LTE are designed to cope with stringent performance levels, i.e. considerable peak data rates and low latencies. With the wide-spreading of LTE networks all over the world, the access to Internet services through mobile connections is increasing considerably, leading to a plethora of mobile services. However, the inherent characteristics of the radio interface results in variable service performance. First, radio resources are shared among the different mobile users and thus the eNB scheduling functions determine the assigned share of resources for each user in a dynamic way. Moreover, the channel quality of each individual user fluctuates over time due to propagation and fading effects. These aspects introduce new issues in the tasks of performance evaluation and adaptive traffic management at different layers of the protocol stack. Currently, there is a gap between advanced research proposals and the actual network and service performance experienced in commercial mobile networks. Although it is generally assumed that the radio channel may introduce considerable quality fluctuations, there is little evidence of the quality dynamics experienced in live LTE networks at different temporal scales. This PhD work aims at providing some insights of these quality variations by analyzing the Channel State Information generated by UEs. Specifically, the document focuses on the downlink direction and on the effects of using spatial multiplexing techniques. From field testing, the main characteristics of the channel quality are studied at the scale of one second. Besides, LTE emulation tests allow evaluating the characteristics of the reported channel quality at the scale of milliseconds. Therefore, valuable information is captured for different layers of the protocol stack. Finally, the impact of the dynamics of these Channel State Information values into different adaptive traffic management schemes is evaluated. For a comprehensive analysis, different mobile network architectures are considered including classical LTE systems and novel proposals such as centralized radio access networks and mobile edge computing.