Bid farewell to "edge", and now edge computing has entered a new channel of rapid development! The industry predicts that from 2018 to 2022, the annual compound growth rate of the relevant market scale of global edge computing will exceed 30%, in which the video business is regarded as the most realistic market demand driving the rapid development of edge computing."Edge computing has entered the 'landing' stage, and video application has become an important 'focus'." Yu Wei, director of the Business Research Institute of China Mobile Research Institute, said in an interview with mobile labs that the video edge computing capability framework built by China Mobile has five characteristics, which can "match" the needs of multiple applications. At the same time, in order to accelerate the implementation of MEC applications, China Mobile anchors the future and works with partners to build a new ecosystem of video edge computing.Accelerated landing, video application is an important "focus" for MEC development5g new infrastructure is now quite hot. With the continuous increase of 5g new construction, edge calculation has also stood at the wind outlet. It can be said that embracing 5g new opportunities and enabling new infrastructure starts from edge computing!Edge computing will also enter the period of large-scale commercial deployment in the next 1-3 years. Gao Tongqing, deputy general manager of China Mobile Communications Group Co., Ltd., pointed out that at present, China is vigorously promoting the construction of new infrastructure, of which 5g is the "first" of new infrastructure, data center is the "base" of new infrastructure, artificial intelligence is the "wisdom" of new infrastructure, and edge computing is the explosive point to realize the integration of the three."With the rapid development of 5g and mobile Internet, the demand of various businesses for edge computing is very urgent. Edge computing can provide value in three aspects." Yu Wei said that first, reduce the consumption of super large traffic or low value traffic on the network and unload the traffic nearby; Second, provide ultra-low time delay for time sensitive applications such as industrial control to realize real-time control; Third, improve the security of private data processing, complete the calculation at the edge, and ensure that the data does not leave the park.Yu Wei pointed out, "at present, there are many typical edge computing examples in the industry. Edge computing solutions are provided for parks, monitoring, ports and other scenes. It can be said that edge computing has developed to the stage of practical application."Although MEC has entered the "landing" stage, it still faces many challenges in terms of standards, application and collaboration.In terms of standards, the traditional core network is still a tight coupling stage, and the key components of edge computing technology need to be decoupled, which requires the openness of standards and specifications;In terms of application, currently, MEC applications focused in the industry are mainly focused on mobile. Many industry applications are originally based on fixed networks. There are still many problems to be discussed on how to realize the integration of fixed and mobile and provide customers with industry upgrading schemes of fixed and mobile integration;In terms of collaboration, some key network elements of the core network extend to the edge. While the implementation of MEC requires a lot of collaboration between edge devices and operators' communication core network, which will bring the demand for cloud edge collaboration. How to meet this demand is still being explored in the industry."From the perspective of industrial development, how to provide open capability services for video and other applications is an important problem faced by MEC development." Yu Wei stressed that MEC development needs to be promoted by large-scale application, and video applications have a large scale in terms of traffic, equipment or application types, which is an important focus of MEC development, Therefore, the design, implementation and opening of common basic video capabilities to better serve all kinds of video applications is an important problem to be solved in the development of MEC.Capability sinking: video edge computing capability framework "matches" multi application requirementsWith the efforts of China Mobile and all parties in the industrial chain, MEC has embarked on a journey of "riding the wind and breaking the waves".MEC 2B business focuses on meeting the requirements of large bandwidth, low delay, security and privacy in the vertical industry, enabling the digital upgrading of enterprises; MEC 2C business focuses on meeting the sinking of "CDN content", local computing and storage, and helping to "re upgrade video consumption". It is reported that China Mobile has promoted about 105 pilot projects in Jiangsu, Zhejiang, Guangdong, Fujian and other provinces, covering 15 subdivided industries."No matter video surveillance, smart city, industrial Internet, life and entertainment, video edge computing is everywhere!" Yu Wei further added: "In the field of video surveillance, application scenarios mainly include face recognition, vehicle recognition, video structuring, etc.; in the field of smart city, application scenarios mainly include smart light poles, smart parks and smart buildings; in the field of industrial Internet, application scenarios mainly include industrial quality inspection, industrial collaborative control, Park logistics, etc.; in the field of life and entertainment, application scenarios mainly include HD video Frequency, smart home security, cloud games, vcdn, etc.; in industrial applications, application scenarios include smart campus, smart enterprise, smart government, etc. ""At the same time, with the advent of 5g big video era, content must sink." Yu Wei pointed out that in order to further boost the development of MEC industry, China Mobile Research Institute has proposed a video edge computing capability framework, which combines cloud video and AI capabilities with network advantages to sink to the edge side and end side, supports the penetration of video capabilities into edge devices, and helps users realize video edge computing intelligent products and applications.It is reported that the video edge computing capability framework has five technical characteristics: according to the application requirements of video edge computing, it can provide basic common video processing functions such as video acquisition, encoding and decoding, and adapt to a variety of scenes; sink video intelligent computing to the edge, support AI model edge reasoning and distributed training, and realize federal learning; adopt virtualization and container technology to manage multiple scenes Three types of video edge devices; support the unified management of the platform of edge devices to reduce the management overhead; compatible with edge devices realized by general-purpose processors and a variety of video intelligent chips to reduce the equipment cost and development cost.It is worth noting that although edge computing drives the integration and development of 5g network and new technologies in various fields, it has realized the further extension of 5g capability. However, in the video category, there are IPTV, surveillance video, CDN of mobile network and other applications. These applications have different angles and different functions. Different requirements are reflected in MEC, which is to implement adaptation schemes in different directions.Yu Wei pointed out that through the video edge computing capability framework, China Mobile Research Institute hopes to solve the problems of intelligent processing, codec and transmission in the video processing of various applications on the edge side, realize the sinking and landing of cloud video AI capability to the edge side, be compatible with and manage multiple types of edge side and end-to-end AI devices, provide video AI processing services online, R & D quality inspection, VR, etc Typical applications will eventually form a closed loop of Pan video access, edge device management, AI capability sinking, end-to-end AI reasoning, edge Federation learning, data characteristics and cloud on model, so as to realize the matching of different application requirements.Anchoring the future: China Mobile joins hands with the industrial chain to build a video edge computing ecosystemIn order to "ride the wind and waves" in edge computing, landing is a key step.Whether operators, Internet companies, cloud computing providers or customers in other industries hope to realize digital transformation and upgrading with 5g and edge computing. Only by working together in the industrial chain can we incubate edge applications with practical value and real returns.For a long time, China Mobile has attached great importance to the development of edge computing, actively promoted collaborative innovation among industry, University, research and application, and continuously promoted industrial development from the aspects of industrial cooperation, standard setting, platform building and application demonstration.In terms of industrial cooperation, China Mobile has established an open laboratory for edge computing and initiated the establishment of an edge computing Committee in China communication society, becoming an important academic and industrial exchange platform in the field of domestic edge computing.In terms of standard formulation, China Mobile takes the lead in the formulation of standards in the field of edge computing and edge cloud among ETSI, CCSA and other domestic and foreign standardization organizations, so as to provide the basis for the industry to form a common technical system.In addition, in terms of platform building, China Mobile actively promotes the research and development of openupf and opensigma edge computing products, builds the service capability of network edge integration, and provides high-quality network edge integration series products for partners.In order to enable MEC to set off a higher industry storm, at the recent "5g e" network edge integration technology summit convened by China Mobile, China Mobile lit up the "100 " node of edge computing. Through strong network, multi computing power, wide distribution, heavy capacity and multi computing power, China mobile provides full stack service capability for "100 " nodes.The above layout has "built the foundation and paved the way" for the development of China Mobile's video edge computing ecology. The industry predicts that the integration of edge computing into video business should be realized on a large scale in the past 12 years. Yu Wei said, "in China Mobile, big video has also become one of the important strategies, and big video technology, capabilities and products are the important development direction of the company."