Queueing Theoretic Models for Multiuser MISO Content-Centric Networks with SDMA, NOMA, OMA and Rate-Splitting Downlink
Multiuser, Multiple Input, Single Output (MU-MISO) systems are proving to be indispensable in the next generation wireless networks such as 5G and 6G. The spatial diversity of MISO systems have been leveraged in physical layer designs in these wireless systems to improve the capacity. Several recent studies have utilised redundancies in the content request along with the spatial diversity of a MISO system to improve the capacity further. It is shown that Max-Min Fair (MMF) Beamforming schemes for MISO based on SDMA, NOMA, OMA and Rate-Splitting could be used to improve the content delivery rates. However, in most of these studies the key aspects such as the queueing delays in the downlink and the user dynamics have generally been ignored. In this work, we study how the interplay between queueing, beamforming and the user dynamics affects the Quality-of-Service (user experienced delay) of downlink in MU-MISO content centric networks (CCNs). We propose queueing theoretic models that are simple in nature and can be directly adapted to MU-MISO CCNs to perform optimal multi-group multicast downlink transmissions. We show that a recently developed Simple Multicast Queue (SMQ) for SISO systems can be directly used for MU-MISO systems and that it provides superior performance due to its always-stable nature. Further, we observe that MMF Beamforming schemes coupled with SMQ can be quite unfair to users with good channels. Thus, we propose an improvement to SMQ called Dual SMQ which addresses this issue. We also provide theoretical analysis of the mean delay experienced by the users in such MU-MISO CCNs.