Guangdong Mobile has worked with ZTE in building large-scale MEC trials in Guangdong Province to support growing service needs in the 5G vertical industries and promote the research and commercial launch of new edge computing/edge cloud projects. They have deployed multi-level MEC in Guangzhou, Jiangmen and Zhuhai and have explored many aspects, involving SA networking and surrounding resources, construction mode, cooperation model with third parties, and edge infrastructure management. These provide valuable experience for 5G construction and future widespread deployment.
Network Performance Verification
Guangdong Mobile's MEC infrastructure contains four levels: city and prefecture core office, important convergence office, ordinary convergence office, and service convergence office (Fig. 1). The MEC environment is built in many offices at different levels, such as Guangzhou and prefecture core office, district and county important convergence office, ordinary convergence office, and campus service convergence office. This helps to gain experience in MEC networking and planning.
The overall project is divided into two phases of verification, including MEC networking deployment, platform function tests, and service trial tests that contain latency test and QoS test.
5G UPF offload capability is the core for operators to develop edge computing/edge cloud, which takes network services as the core to drive edge cloud services and thus to achieve network-cloud convergence. Therefore Guangdong Mobile has implemented in the first phase a pre-commercial validation of the MEC environment including SA networking and surrounding resources, construction mode, cooperation model with third parties, and edge infrastructure management. Both general-purpose servers and special hardware resources such as integrated server, smart NIC, and GPU are deployed on each edge computing node to form diverse hardware resource pools that can meet different needs and office conditions. This results in a heterogeneous unified resource pool management capability. A lightweight platform is used at the edge, which provides three types of resource environments including virtual machine (VM), container, and bare metal to meet the needs of different edge service applications and deployments. An MEPM is centrally deployed in the core office in Guangzhou to perform unified operation and management of MEPs at each edge node. A unified cloud management is also deployed to manage heterogeneous edge resource pools. Portal provides services in the multi-tenant mode, and third-party APPs are deployed on VMs or containers provided by edge nodes through the self-service portal.
Business Model Verification
Industry digitalization is the focus of the second half of the internet. Vertical industries, cloud providers, and operators all take edge computing as the focus of their network deployment. Guangdong Mobile arranged MEC trials in the second phase to verify MEC IaaS/PaaS/SaaS three-layer solution that supports the converged deployment of third-party and self-owned applications under the telecom cloud architecture, and verify local traffic offload and interface capability exposure. By combining network slicing, Guangdong Mobile also carried out trial verification and exploration of end-to-end technical architecture, service quality assurance capability, and business model. For 2B and 2B2C scenarios, Guandong Mobile has introduced third-party partners involving various service applications such as CDN, cloud games, face recognition, video live broadcast, DPI, V2X, smart port, smart forklift, and VR. Diverse products and services are provided to meet different needs of industry users.
—It provides game rendering and acceleration service in the important convergence office for Tencent cloud games, delivering low-latency game experience through MEC (with an end-to-end latency less than 15 ms).
—It deploys V2X service on the MEC of the vehicle industry park, helping the vehicle research institute carry out research on IoV applications such as 5G driverless car and remote driving.
—It deploys image recognition service on the MEC, exploring future intelligent video surveillance services oriented to specific scenarios.
—It provides the port with a private 5G campus network that features data security, low latency, and high reliability. Through the nearby MEC processing, the network carries out video monitoring and remote control of quayside cranes in the port area.
—It provides video entertainment enterprises with professional acceleration and video transcoding capabilities, meeting high-bandwidth and large-storage requirements for CDN and VR VOD.
According to service scenarios and latency requirements, MEC can be flexibly deployed in the service convergence office, ordinary convergence office, important convergence office, and city and prefecture core office to interconnect with Guangzhou Mobile's 5G core network. Through the MEC trials, the system provides flexible and packable modes for different application scenarios, and fully verifies edge infrastructure, edge office, basic edge computing, and general edge service capabilities. By reducing costs through the large-scale replication, Guangdong Mobile can achieve a balance between construction costs and service needs.
User Experience Enhancement
Through the implementation of the project, network requirements such as SPN/PTN, IP bearer and CMNET under different office conditions have been fully verified, and the networking capability has been improved. Third-party applications have also been fully verified in various scenarios and modes, such as high-performance data forwarding, unified management of edge resource pools, applications of typical configuration models such as UPF/MEP, differentiated service guarantee of E2E slice, as well as slicing mall application signing, ordering, and multi-dimensional billing. These will provide valuable experience for large-scale commercial deployment of 5G MEC in the future.
Operators can seize the opportunities of 5G development, leverage their network advantages (such as network capabilities, users, traffic, and data) and infrastructure advantages (such as offices, transmission, and platforms), and focus on vertical industries and OTT applications. By actively building differentiated edge computing platforms and service capabilities, they can achieve the cloud-network-edge-terminal synergy, thus greatly reducing latency, improving user experience, and transforming their networks from access channels to information service enabling platforms.