Network solutions for CoMP Coordinated Scheduling

Demanding throughput, latency and scalability requirements of 5G networks may be addressed by relying on dense deployments of small cells. Coordinated Multipoint (CoMP) Coordinated Scheduling (CS) techniques are introduced to reduce inter-cell interference in case of dense deployment, given that local CoMP-CS information from the evolved NodeBs (eNodeBs) in the cluster are timely collected at the scheduling decision entity. This work studies how the distribution of CoMP-CS cell information is affected by the backhaul infrastructure in terms of both physical and logical topology. The differentiation between physical and logical infrastructure is justified in the context of new approaches like Software Defined Networking and Network Function Virtualization that enable the dynamic configuration of the network. We consider either a Packet Switched Network with three possible topologies (namely, ring, mesh and star) or a Time Division Multiplexing Passive Optical Network (TDM-PON), both carrying heterogeneous traffic. To improve the convergence time in the TDM-PON, we propose a novel bandwidth allocation scheme to prioritize the signaling traffic with respect to data traffic. Performance of both distributed and centralized CoMP-CS are compared in terms of convergence delay and traffic overhead. Finally, we analyze the impact of the periodicity of CS operations on mobile performance, in terms of average UEs throughput, in the presence of different cell loads.