High Throughput and Low Complexity Traffic Splitting Mechanism for 5G Non-Stand Alone Dual Connectivity Transmission

Jiayun Sun, Shunqing Zhang*, Shugong Xu, Shan Cao

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

5 Citations (Scopus)

Abstract

Dual connectivity (DC) is a key technology to improve the throughput of the fifth generation communication networks (5G) system. Due to the limitation of the splitting strategy and transmission delay of different transmission nodes, the out-of-order phenomenon of the packet data convergence protocol (PDCP) layer is bound to occur in the actual DC system. Packet out-of-order in PDCP layer will affect the end-to-end throughput of PDCP layer and there is no effective solution at present. Existing schemes either neglected to consider the transmission rate difference between New Radio (NR) Base Station (BS) and Long Term Evolution-Advanced (LTE-A) BS, or performed unified modeling for queuing dynamics, or did not consider the existence of Xn delay, thus failing to model the out-of-order problem of PDPC layer. In this paper, we modeling PDCP out-of-order into a hierarchical model and continuously reduce the maximum out-of-order depth through recursion. Two traffic splitting mechanisms are proposed based on user equipment (UE) feedback and BS observation. These two low complexity traffic splitting mechanisms both show efficient link resource utilization ratio of more than 90% in static and mobile user scenarios, and are superior to conventional algorithms.

Original languageEnglish
Article number9417236
Pages (from-to)65162-65172
Number of pages11
JournalIEEE Access
Volume9
DOIs
Publication statusPublished - 2021
Externally publishedYes

Keywords

  • 5G
  • dual connectivity
  • low complexity
  • out-of-order
  • throughput
  • traffic splitting

Fingerprint

Dive into the research topics of 'High Throughput and Low Complexity Traffic Splitting Mechanism for 5G Non-Stand Alone Dual Connectivity Transmission'. Together they form a unique fingerprint.

Cite this