多接入边缘计算在智慧矿山网络中应用分析

Application analysis of multi-access edge computing in intelligent mine network

  • 摘要: 针对当前矿山网络软硬件设备种类繁多、制式复杂、资源共享率低,网络升级改造困难,因上下行传输带宽受限、服务总时延过长而导致垂直业务需求无法得到满足等问题,提出在矿山网络中引入多接入边缘计算(MEC)技术,为实现智慧矿山提供支撑。针对智慧矿山视频监控类、时延敏感类、连接类业务,探讨了MEC应用模式,分析了MEC支持多种协议和设备接入特性,指出MEC是基于现有井下多种协议和设备提供融合管控的唯一选择,是为用户提供端到端、无差别高质量服务的最佳解决方案。建立了基于业务需求的矿山网络计算树模型,将其演进发展划分为Pre−5G、5G、Beyond−5G(6G) 3个阶段,分析了各阶段的主要特征和标志性技术,以及负载均衡性、业务移动性、数据安全性、可扩展性、一体化特性等业务特性,指出各特性在第1阶段最差,在第3阶段最优,但距离实现硬件设备真正通用、接口完全标准化的目标还存在差距。结合计算树模型,提出了一种智慧矿山网络MEC部署模型,指出该模型基于深度为2的三级扁平化结构,能够充分发挥各级节点的边缘计算能力,在兼顾各节点负载均衡的同时动态选择任务卸载策略,从而完成高效任务协同,且各节点功能可根据地面/井下实际情况按需取舍、灵活调配,适用于智慧矿山建设的各个阶段。

     

    Abstract: In order to solve the problems of various types, complex system and low resource sharing rate of software and hardware equipment, difficult network upgrading and transformation, and the incapability to meet the requirements of vertical services due to limited uplink and downlink transmission bandwidth and long total service delay, this paper proposes to introduce multi-access edge computing(MEC) technology into mine network to provide support for the realization of intelligent mine. For the intelligent mine service types of video monitoring, delay sensitive and connection, the application mode of MEC is discussed, and the characteristics of MEC supporting multiple protocols and equipment access are analyzed. It is pointed out that MEC is the only choice to provide integrated management and control based on the existing underground multiple protocols and equipment, and is the best solution to provide undifferentiated high-quality end-to-end service for users. The intelligent mine network computing tree model based on business requirements is established, and its evolution is divided into three stages, namely Pre-5G, 5G and Beyond-5G(6G). The main characteristics and symbolic technologies of each stage are analyzed, as well as the business characteristics such as load balance, business mobility, data security, scalability and integration characteristics. It is pointed out that each characteristic is the worst in the first stage and the best in the third stage. However, there is still a gap to achieve the goal of truly universal hardware devices and fully standardized interfaces. Combined with the computing tree model, the intelligent mine network MEC deployment model is proposed. It is pointed out that the model is based on a three-level flat structure with depth of 2, which can give full play to the edge computing capability of nodes at all levels, and dynamically select task unloading strategy while taking into account the load balance of nodes, so as to complete efficient task coordination. Moreover, the functions of nodes can be selected and flexibly deployed according to the actual situation on the ground or underground, which is suitable for all stages of intelligent mine construction.

     

/

返回文章
返回