Abstract: Edge computing, as the evolution and development of cloud computing technology, has the characteristics of close application, intelligent and flexible, and efficient distribution. Therefore, cloud computing/edge computing technology has a wide application prospect in the Internet of Things, industrial manufacturing, transportation and other fields. Combined with the characteristics of edge computing and manufacturing industry, the application of edge computing in intelligent manufacturing is emphasized. Firstly, four key factors and capabilities of edge computing application in the industry are analyzed, including 5G technology, virtualization, flexible deployment and industry attributes. Then, from three levels of operation, market and technology, the demand, challenge and application opportunity of intelligent manufacturing field are analyzed in detail. Finally, the application modes of edge computing in intelligent manufacturing are given, including management mode, business mode and deployment mode, and the application direction of edge computing in intelligent manufacturing is preliminarily studied.
01 before it
Edge computing (EC) refers to the computing environment with the most advanced telecom network attributes deployed at the edge of the network (i.e., edge computing network or edge cloud). It is a system of “hardware + software + services”, which is different from the embedded edge computing capability that sinks computing to the device side. Reduce latency for network operations and service delivery by providing IT service environments and cloud computing capabilities at the edge of the network. Its technical characteristics mainly include proximity, low latency, high broadband and location cognition, etc. It has broad application prospects in the future, such as intelligent manufacturing (industry 4.0), Internet of vehicles (unmanned driving), AR, video optimization acceleration, surveillance video analysis, etc.
Although network virtualization is not the “originator” of edge computing and network slicing, as ICT (Information and Communications Technology) convergence technology and ecology gradually mature, virtualization architecture is becoming a reality for commercial deployment step by step. The full virtualization and complete telecommunication network element functions carried by it and the inherent flexible distributed deployment mode formally trigger the mutual help and complement of edge computing and network slice. The combination of the two provides core capabilities for the telecom network in the industrial ecological construction of 5G.
The target market of edge computing has changed from the traditional 2C to the new blue ocean dominated by 2B. This new market is also the field that the original public cloud/cloud computing platform has not successfully entered, which is related to the technical strategy positioning of public cloud itself. At the same time, it is also the typical pain point of the industry market due to the lack of coverage near the edge of the enterprise.
Edge network computing and complementary relationship is shown in figure 1, edge computation slicing through the network to the industry of private network across geographical coverage and diversity of the network resources and provides the guarantee performance, and network section is a support at the edge of the deployment of computing environment, makes the network capacity in the industry which fall to the ground.
02 Edge computing industry key capabilities
If using edge positioning calculation to create valuable products for the manufacturing sector, and realize intelligent manufacturing and industrial transformation of 4.0, so the wise manufacturing scenario, edge calculation in addition to the congenital advantage of the cloud computing, also must have some key abilities and characteristics, the ICT maintain strong competitiveness and vitality in the ecological environment.
(1) 5G wireless network
As carriers of core business, but also its own core competitiveness, cellular wireless network technology is the edge on the basis of the authorized spectrum calculation should take advantage of the most basic and the most powerful ability, especially the 5 g technology in the improvement of network bandwidth and transmission delay leap type has become an important key operators to enter the industry market.
(2) Virtualizing telecommunication networks
Virtualization technology is the key product of telecom network embracing IT technology, and is also the representative technology of advanced productivity in the era of cloud computing. Although NFV (Network Function Virtualization) does not have obvious technical advantages compared with cloud computing in terms of underlying technology implementation, However, through the support of telecom network capacity, the virtualized telecom network can bring another unique differentiated competitiveness for edge computing services, that is, the opening of network capacity to edge services.
(3) Flexible and fast deployment
Industry market fragmentation is serious, and the ecological environment of each industry has its own history and complicated competition pattern, the enterprise size is differ,’s ability to accept new technology is different, the telecom operators as a newcomer to the field, although back against the powerful resources and technology, including policy support, But the beginning must be a period of trial and error. Based on this understanding, the deployment capability of edge computing network must have these five characteristics.
• Lightweight network, avoiding heavy and complex existing telecom network architecture.
• Flexible networking and on-demand networking do not force carriers and enterprises to have unique boundaries on network environments and assets.
• Open architecture, the pursuit of a general IT platform and IT integration interface, to ensure maximum compatibility with the original industry technology environment.
• Rapid deployment and the impact on the industry’s production environment must be one of the important indicators of the industry’s acceptance of edge computing.
• Easy to manage, the introduction of edge cloud should not bring disproportionate management difficulty and cost increase to the enterprise.
(4) Industry capability attributes
Since edge computing is targeted at the industry market, the integration of industry understanding and expertise is a reflection of the differentiated value that operators can provide. In this regard, operators can extract valuable experience from their own network construction and systematic network construction and business development experience accumulated in the process of network evolution, and apply it in the innovation and transformation of the industry to help enterprises get a head start in their own market competition.
In short, the target market of edge computing network must be a new blue ocean market facing vertical industry, which is different from the mature market with public cloud as the main business form and 2C as the main application scenario. In the era of 2G/3G/ 4G telecom network and cloud computing, this field has not been fully developed due to limitations in data security, business control, network performance and other aspects, which are closely related to the four elements mentioned above. Therefore, the key to the success of edge computing in 5G is how to make the industry accept and adopt the technical characteristics brought by these elements to the maximum extent and then transform them into the improvement of its own productivity.
Relying on the most advanced 5G cellular wireless transmission technology and supported by the virtualization telecom network architecture, edge computing provides the most flexible network coverage for the society and industry. Meanwhile, it can unload traffic pressure for the backbone network through data diversion and effectively save the overall network investment in the later stage. On this basis, the maximization of network value is pursued by opening up and providing computing resources and business capabilities with industry attributes, creating new commercial value and market opportunities for the industry and operators themselves. Therefore, the four elements help and complement each other to form a joint force, and computing at the edge of 5G brings another great opportunity for vertical industry to leap forward.
Demand analysis of edge computing intelligent manufacturing application
The conditions of large-scale deployment of edge computing network involve many aspects such as technology, market and operation, and the effect of its industrial application directly depends on the maturity of relevant aspects. According to the above industry application analysis, the application research of edge computing in the field of intelligent manufacturing must first clarify the problems, challenges and opportunities in the development of the industry.
3.1 Operation level: Clear target industry and clear idea of operation mode
Edge computing, the deployment form of telecom network closest to the enterprise environment, brings multiple possibilities for cooperation among enterprises, operators and solution providers. The pressure of the vast majority of enterprises because authors efficiency makes it accept the latest Internet technology at the same time must also be considered acceptable to the input-output ratio, and 5 g and edge calculation adopted for most of the traditional enterprise is still green, so the telecom operators and manufacturers need to play a more active evangelist, play a vanguard role.
3.2 Target market level: the enterprise’s own preparation and acceptance of new technology
Edge computing is a technological innovation for most traditional enterprises, and the technological preparation of enterprises is also an important condition, including the status quo and acceptance of enterprises themselves. In addition, large-scale deployments can be analyzed from these perspectives.
(1) The breadth of industry coverage
The deployment of edge computing network in the initial stage, what direction to take the first step, how much steps to take, which industries to target, and how to plan the scale of network construction and coverage are currently the focus of discussion between operators and core manufacturers.
(2) The richness of application scenario types
The application scenarios that edge computing can cover are very rich. Based on the analysis and summary of more than 200 use cases in ten global industries, it can be divided into the following 9 categories of typical scenarios (as shown in FIG. 2) from the perspective of technical scenarios, which are related to multiple specific industrial fields, and some individual industrial fields contain multiple technical scenarios.
Each scenario puts forward strict requirements for network connection, network architecture and software openness to edge computing network. Every commercial solution with practical capability has many tasks to be verified and promoted by all parties in terms of time to market and requirements for network deployment.
3.3 Technical level: perfect edge network technical framework and commercialized edge computing business platform solution are required
(1) In terms of network bearer, as one of the key functions of edge computing network networking, the implementation mechanism of detailed functions of local shunting technology is still in the process of discussion and development. Before these capabilities are fully standardized and verified, cross-regional edge computing collaboration is still facing challenges.
(2) In terms of support for business scenarios, the planned technical architecture of edge computing network should be able to seamlessly integrate with typical scenarios of industrial applications, which may be divided into two categories: The first is the business sensitive to the wireless network environment, such as the need to rely on the high bandwidth characteristics of 5G to upload the business data to the computing environment of the edge cloud and ensure low delay to send the calculation results to the receiver, such as the robot cloud in the intelligent factory or remote control and many typical Wi-Fi alternatives. Remote surgery, parallel driving in V2X and autonomous vehicle applications; The other category is scenarios that do not rely on the latest wireless transmission technology, but have hard requirements for cloud and centralized deployment of computing resources, such as terminal entertainment services (online games /AR/VR) that hope to reduce the cost of terminal equipment (or not increase the cost) while improving user experience.
(3) Edge computing network is determined by the target market, destined to win the acceptance and popularity of enterprise customers with lightweight network architecture and management mode, and this step is the time to give full play to the achievements of network virtualization of operators, of course, but also the practical assessment and verification of NFV architecture.
(4) The large-scale deployment of edge computing is inseparable from a complete set of operation management platforms with basic functions and corresponding management mechanisms. From the fierce competition between ETSI MEC and 3GPP at the beginning, to today’s PaaS construction plan based on the actual network conditions and market environment in China, combined with operators’ own positioning objectives, Whether telecom operators can open the construction of edge computing from the ecological needs to wait and see.
(5) In addition, the maturity of the end modules and when to start mass supplying them to the market are also key factors. Many industry applications that are strongly related to edge computing also have a strong dependence on the integrity of 5G terminals/modules to achieve relevant 5G functions.
Application research of edge computing intelligent Manufacturing
4.1 Analysis of intelligent manufacturing concerns
As a new entrant, operators and equipment providers must have a full understanding of the current situation of the industry they are entering. If the 5G technology and edge computing solutions brought by them can meet the actual pain points of enterprises, they will greatly accelerate the implementation of the new market and thus achieve the purpose of successfully entering the new market. However, due to differences in various industries, it is necessary to have a basic plan for the target market, invest resources in specific industry fields, and excavate the best entry point by means of in-depth field investigation and analysis inside enterprises.
This paper summarizes the typical pain points of the manufacturing industry, as shown in Figure 3. Other industries have their own pain points and should be singled out as an area of particular concern for edge computing deployment.
In other words, it is inevitable that 5G will continue to be piped if it fails to come up with technologies and solutions that target specific pain points of enterprises and stays in the role of a provider of basic universal network connectivity capabilities.
4.2 Research on deployment forms of intelligent manufacturing
As the second section puts forward “flexible rapid deployment form”, in order to adapt to different industry on the edge of computing and 5 g ability of differentiation, edge cloud to the ground in the business there are a variety of deployment form, each should eventually fall to the ground network network architecture design and business requirements, management mode, operation mode and commercial mode. The following three types of typical network deployment patterns are summarized, as shown in Figure 4.
(1) Enterprise private network
Complete telecom network deployments provide end-to-end network coverage in an enterprise environment. This ensures the isolation and independence of enterprise data and network environment to the maximum extent, but also puts forward high requirements on enterprise network management capability.
(2) Enterprise private network
The user plane gateway is deployed near the enterprise site, and the other nes of the core network are deployed on the upper nodes of the telecommunication network. This mode can not only maintain the unified management of the network by operators, but also meet the security demands of data not leaving the factory due to the sinking of the user surface. In particular, enterprises with limited technical support capability will not have high maintenance and management technology requirements.
(3) Network slicing networking
Network slicing in 5G architecture. Of course, this mode requires operators to judge the target market and plan the deployment of corresponding network resources, especially for the distribution and scale of edge computing nodes, which is also the focus of the joint efforts of equipment vendors and operators.
4.3 Research on intelligent manufacturing management mode
At present, the focus of the edge computing node management mode is whether to bring the management of the edge cloud into the scope of unified NFV business orchestration management, or to maintain a relatively independent management system for it. According to the possible deployment modes of edge cloud described above, if the service domain is separated from the network domain, an independent service domain management system is more likely to be established for edge cloud, while the network domain is merged into the unified management domain or independently deployed locally depending on the three typical network deployment modes.
• Give full consideration to IT level and management mode of customers in different industries and management demands of builders of edge nodes, and provide hierarchical operation management mode.
• For different industrial application requirements, start with application scenarios that do not have high requirements for SERVICE-level agreements (SLA), and gradually enter key application areas after the technology and cooperation are mature.
In short, the management mode of edge nodes should provide necessary and reasonable management and operation capability realization for operators’ edge network capability realization under the principle of “not bringing disproportionate management difficulty and cost increase to enterprises”.
4.4 Research on intelligent manufacturing business model
At present, in the early stage of network construction, it is really not appropriate to determine the business model prematurely, and there will not be only one model in the future. It is suggested to keep an open mind in the early stage and explore multiple models in parallel. Based on the three types of networking modes proposed above, the possible business models are simply analyzed, as shown in Table 1.
The business model proposed above can be used as an exploration model in the initial stage of edge computing deployment. However, with the deepening understanding of the industry market and the deepening understanding of edge computing and 5G technology, more and more reasonable business models will be found.
05 conclusion
5 g technology and the edge of the network of telecom business in vast and carve out a piece of brand new blue ocean industry market, as a new entrants need to offload lightly, don’t have the idea of traditional telecom business, adjust the attitude to penetrate deeper into 2 b, in the process industry must identify pain points, the integration of its own technology and resource advantages, Keep an open mind and actively seek win-win cooperation, and strive to find pioneer enterprises and industry partners who eagerly embrace the most advanced network technologies with the help of digital transformation. Only in this way can pipelinization be avoided, 5G technology and edge computing be brought into full play, and the top of the value chain be constantly approached in the new ecological construction.
References:
[1] 3GPP. Enhancement of Support for Edge Computing in 5G CoreNetwork (5GC): TR.23.748[S].2019.
[2]3GPP. Application Architecture for Enabling Edge Applications: TR.23.758[S].2019.
[3] ETSI. MEC in an enterprise setting: white paper No.30[S].2018.
[4] ETSI.MEC in 5G networks : white paper No.28[S].2018.
As a person
Kai CAI (1972−), male, senior engineer of China Mobile Chongqing Co., LTD., member of Chongqing Communication Industry Expert Committee, mainly engaged in the research of core network IMS, virtualization technology and 5G edge technology.
CAI Zhengyun (1979−), male, engineer of China Mobile Chongqing Co., LTD., chongqing electromagnetic radiation environmental assessment expert, mainly engaged in the research work of core network, wireless network, and wired network 5G virtualization.
Li Yu (1984−), male, currently works for China Mobile Chongqing Co., LTD., mainly engaged in the research of core network IMS, virtualization technology and 5G edge technology.
Fu You (1988−), male, engineer and core network operation and maintenance supervisor of China Mobile Chongqing Co., LTD., mainly engaged in research work on core network and 5G virtualization technology.