How can edge computing win in the context of 5G "new infrastructure"?

Edge computing has become one of the hottest technologies in the development of 5G. The Empowered Edge, one of Gartner's top ten technology trends in 2020, also demonstrates the importance of edge computing.

[[328319]]

At present, the construction of new infrastructure (referred to as "new infrastructure") is in full swing, and edge computing has naturally ushered in a new trend. Different fields, different perspectives, and different technical systems have come together for the goal of edge computing. In the cloud computing market, there are edge clouds and telecom clouds; in the field of industrial Internet, AII and ECC Industry Alliance have established an edge computing working group to share results; in the construction of operators' 5G networks, MEC has begun to move towards openness and comprehensive decoupling, from OT to IT, then from IT to CT, and finally to OICT.

Market cultivation requires patience, and operators face difficult choices

Conceptually, edge computing includes three stages: connection, intelligence, and autonomy. Currently, most applications are still solving the problem of connection, so it is difficult to show the unique value of edge computing. The real value of edge computing lies in intelligence and autonomy. Intelligence provides computing power while reducing latency, while autonomy increases reliability while reducing costs.

End users will not care whether operators use edge computing or centralized computing. Therefore, operators can treat edge computing as cloud and sell resources; treat network capabilities in edge computing as services and sell services; transplant vertical industry applications to the edge and sell applications. However, this model will not be established overnight, and market cultivation still requires patience.

From the perspective of operator construction, the attitude is firm, because edge computing is the link between 5G construction and applications. Cultivating user habits and building an ecosystem all depend on the speed of application mining and market cultivation. Judging from the pace of operator investment, investment in edge computing is cautious. In 2018, operators sorted out nearly 10,000 computer rooms that meet the conditions for edge computing deployment, but there are not many that are truly large-scale, mainly because the applications are not heavyweight and large-scale enough.

It is for this reason that operators have been making choices. If they do not leverage the advantages of mobile access, that is, use access rooms and network edges as sinking nodes for cloud computing capabilities to provide services, they will lose many business opportunities and become pipeline providers. It will be extremely difficult for operators to quickly recoup their investment in 5G with such a large investment.

Edge computing can reduce costs, but technology homogeneity is serious

You must know how to calculate when doing edge computing. A customer who has demand for computing power, latency, and bandwidth must consider whether it can save costs, otherwise no user will pay for it. Of course, after edge computing develops on a large scale, the cost will be reduced, and finding highly replicable scenarios will become a top priority. Among them, in projects such as smart cities, smart security, and smart factories, edge computing can reduce backhaul bandwidth, link leasing costs, storage costs, and management costs. Therefore, in the field of infrastructure construction with large investments, the cost savings of edge computing may become an important reason for everyone to choose it.

From a technical point of view, the edge computing framework is nothing more than the ECC Industry Alliance's Edge Computing 3.0 Framework and ETSI's reference framework for MEC. Operators are moving towards an actively open architecture. China Unicom's white box UPF and China Mobile's OpenUPF plan are both illustrating a trend - the more open, the more value-added. Only openness can attract more players and enable users to enclose and transform. With more and more ecological partners joining, the previous industry barriers will no longer exist as the ecological partners' technology converges, and industry knowledge will become a service that penetrates into various edge applications.

With the cultivation of ecology and market and the expansion of construction scale, the number of edge computing deployment nodes will increase, the technical complexity will greatly increase, and the technical system supporting edge computing will become wider and wider. SDN, NFV, cloud computing, AI, containers and other technologies are likely to become powerful tools for the development of edge computing. The depth of industry knowledge accumulation, the maturity of the technical system, and the breadth and depth of ecological construction will surely become important chips for the success of edge computing.