5G切片架构下具有重传机制的轮询系统研究

面对海量应用场景及客户需求，传统的轮询系统方案不能满足轮询系统中每个站点所需要的服务质量要求。为降低系统时延，适应更大的网络规模，提出基于5G网络切片的轮询方案。将5G网络切片划分为增强型移动带宽切片、海量机器通信切片和高可靠低时延通信切片，再将这3类切片各自划分出N个更低层子切片，构建3个独立的具有重传机制的完全服务轮询模型。信息分组按照先进先出的规则进入各子切片，由通用服务器进行统一发送。在此基础上，采用概率母函数及马尔可夫链建立非理想信道环境下基于5G网络切片架构的轮询系统数学模型，推导系统平均排队队长和平均等待时间的精确表达式，进行信息分组出错概率和重传阈值与轮询系统性能之间的定量关系分析。基于Matlab的仿真实验结果证明了该模型的正确性，其较重传门限服务时延更少，能够为非理想信道环境下基于5G网络切片架构的轮询系统分析提供一种快速评估机制。

Research on Polling System with Retransmission Mechanism in 5G Slicing Architecture

Facing massive application scenarios and customer needs, traditional polling system solutions cannot meet the Quality of Service(QoS) required by each site in the polling system.A polling scheme based on 5G Network Slicing(NS) is proposed to reduce system latency, to enable adaptation to a larger network scale, and meet the needs of different application scenarios.The 5G NSs are divided into enhanced Mobile Broadband(eMBB), massive Machine Type of Communication(mMTC), and ultra-Reliable and Low-Latency Communication(uRLLC) slices, and each of these three types of slices is further divided into N lower layer subslices to construct three independent exhaustive service polling models with failure retry mechanisms.The message grouping enters each subslice according to the first-in-first-out rule, and each subslice is sent uniformly by the universal server.On this basis, exact expressions for the average queue length and average waiting time of the system are derived, using Markov chain and probability mother function, to establish a quantitative relationship analysis between the error probability and retransmission threshold of message grouping while realizing an acceptable performance for the polling system.The accuracy of the model is proven by simulation experiments.This model has a lower latency than the retransmission-gated service.The analysis model provides a fast evaluation mechanism for the analysis of polling systems based on 5G NS architecture in a non-ideal channel environment.