QoS-Forecasting-Based Intelligent Flow-Control Scheme for Multi-Connectivity in 5G Heterogeneous Networks

Abstract

In the fifth-generation (5G) wireless-network system, the convergence of multiple networks of different standards as well as that of high-and low-frequency networks exists since a long time. Owing to the inability of 5G networks to predict the user quality of service (QoS) accurately, these networks are prone to issues such as access congestion, low QoS, and frequent congestion in one network while other network resources remain idle. Therefore, 5G networks fail to meet the QoS requirements and also prevent effective resource utilization. The deployment of multi-connectivity technologies can facilitate the optimization of the multinetwork convergence-system architecture. However, such technologies are faced with several challenges. Existing literature mainly focuses on the development of a multi-connectivity flowcontrol scheme to determine the secondary nodes (SNs) to which the master node (MN) should distribute data. This paper presents a three-step, QoS-forecasting, intelligent flow-control scheme, wherein the user equipment (UE) determines the data-flow direction based on the network characteristics and load handled by each node. Subsequently, the MN determines the initial user priority based on load balancing, user characteristics, and fairness. Finally, the MN allocates data to each SN in accordance with the QoS and average transmission capability of UE. The simulation results reveal that the proposed algorithm improves the system throughput significantly compared to the single connectivity and traditional fixed-data-split methods. Furthermore, the proposed method facilitates transmission-delay reduction, radio-link failureprobability control, and improved system robustness.

Multi-Connectivity Between Terrestrial and Non-Terrestrial MIMO Systems

“…By applying dynamic control algorithms [24], the benefits of multi-connectivity can be reaped in device-to-device (D2D) cellular links [25], and multi-user scenarios [26]. Moreover, multi-connectivity has been widely explored in the upper layer of computer communications in terms of functional architecture design [27], flow-control forecasting [28], userplane application [29] etc.…”

“…where we assume that there is at least one BS in the serving area of a satellite A N , and E Ω N i is given in (28). The probability of case 2 is given by…”

QoS-Aware Downlink Traffic Scheduling for Cellular Networks with Dual Connectivity

“…After collecting UE reports, the MeNB uses predefined functions to make traffic splitting decisions. In reference [5], each UE first decides which SeNB it wants to connect to. After confirming the UE's choice, the MeNB dispatches data to the corresponding SeNB according to the network load and fairness.…”

“…In this situation, if the MeNB dispatches more downlink data to the SeNB, the SeNB needs to spend more radio resources to process the UE u's data, and thus, the system throughput will decrease. Some previous works (e.g., [4,5]) discussed data dispatching mechanisms for the MeNB. However, we found that the previous schemes do not consider SeNBs' capabilities and statuses when making dispatching decisions, and thus, the packet drop ratio increases because data dispatches to SeNBs may not be processed in time.…”