Impact of Link Heterogeneity and Link Correlation on Multi-Connectivity Scheduling Schemes for Reliable Low-Latency Communication

Abstract

Industrial applications like closed-loop control or mobile robots pose strict latency and reliability requirements which cannot all be fulfilled by existing wireless communication systems. Multi-Connectivity (MC), i.e. using multiple communication paths at once, can be a measure to enhance latency and reliability of wireless communication systems. Different scheduling schemes, i.e. Load Balancing (LB), Packet Duplication (PD) or Packet Splitting (PS) can be utilized to distribute packets over the available links. The impact of link heterogeneity and correlation of latency among the links on the MC scheduling schemes needs to be investigated to evaluate the suitability of MC for reliable low-latency communications. In this paper, measurements of currently available technologies, i.e. small cell LTE and WiFi networks, are conducted. These measured traces are then fit to statistical distributions in order to be able to emulate different MC scheduling schemes over links with certain correlation and homo- or heterogeneity. Our evaluation reveals that homogeneous links are beneficial for all three scheduling schemes. The correlation of latency in time among the utilized links shows to have a strong impact on the performance of the different MC scheduling schemes. For uncorrelated links, PD can improve mean and tail latency and increase reliability. For correlated links on the other hand, PD has no gain compared to the single links, while PS can improve mean and tail latency performance. The number of links needed to fulfill latency and reliability requirements of industrial use cases is analytically derived for MC with PD over uncorrelated links.