Analysis of wireless edge networks in the frame of stochastic geometry

Abstract

In this dissertation, we propose an appropriate coded content storage technique and an appropriate coded task offloading technique based on stochastic geometry framework in wireless communication systems. The MDS coding method is known to be the desirable code for retrieving content and obtaining task output result in an wireless edge network, because content and task result can be retrieved by collecting only a given number of MDS coded segments even if the divided part is exactly not known. However, it is not known which MDS encoding method is optimal for content recovery and task retrieval in a wireless random network environment where the location and number of nearby helpers are not given, so this problem is solved in this thesis. In addition, as the introduction of flying base stations is emphasized by coverage expansion, we design a correlated network topology and analyze the average coverage probability as the serving Flying Base Station(FBS) moves over time, which is differently determined how the number of FBSs and the receive beam-width are determined. Through analysis of this, we present guidelines on which non-terrestrial networks are ideal to build, considering the wireless environment and the mobility of FBSs.