Gartner recently released “Edge of Things industrial computing market Guide”, analyzes the industry IOT edge computing market trends. According to Gartner predicts that by 2022, over 50% of businesses to generate data will be created and processed outside the data center or cloud, 20% of new industrial control system will have the analysis and reasoning edge AI, at least 50% of the site was the project will use the container network edge application life cycle management, in addition to CONTROL ENGINEERING China Copyright , at least 50% of the high-end industrial IOT gateway provides an optional module 5G. IIOT edge defined edge computation computing market solutions can transmit data to the data processing and storage in a local or central data center before the cloud. Typical data sources include industrial composition networked environment sensors and control devices, such as programmable logic controllers and distributed control systems. Currently edge computing solutions suppliers from diverse backgrounds, including data center OEMs, OT, cloud computing, analysis and communication service providers. IIOT edge computation need to have the ability to aggregate data key generated by the terminal device and standardized, so that the data analysis platform can receive the data. Data processing capability, for example, by a rule engine event filter, process analysis or complex event stream processing data generation. AI reasoning model and analysis can be executed. Incoming event data and the ability to take local action-based. The ability to provide local and remote visualization capabilities. Ability to transmit and receive data, and the ability to customize the operation of the network interruption between the cloud or any data center. IIOT edge computation deployment pattern edge IOT industrial solutions are often calculated by the edge devices, gateways, I / O modules, edge servers, data centers and micro-analysis software. The speed and the actual use cases need to acquire and analyze data, different combinations of these components can be deployed at the edge. It may be internal to the device, or by internal gateway servers deployed in the edge calculation embodiment, according to the edge, in order to meet different levels of analysis and local decision making needs. The following table depending on the desired level of computing power, the different edges include delivery model calculation key property. Things to deploy edge industrial computing solutions have two models: Equipment – Gateway – Cloud; devices – Gateway – Server – Cloud.According to the type of analysis you need to run in the vicinity of the data source, you can deploy one of these two architectures them. IIOT edge computing the current market trend of things most vendors in the market have been recognized edge computing has become an integral part of the Internet of Things solutions. Due to the high cost of bandwidth-intensive, does not affect the performance or practical reasons, not all data needs to be sent to the cloud or data center core. Therefore, you must deploy data aggregation and processing functions in data generation, the use of real-time analysis to enable rapid decision-making. In some cases, in order to meet regulatory requirements, the edge position needs to be processed and stored data. Gartner expects the industry Internet of Things will go through a period of consolidation. OT major supplier of industrial platform vendors things, analysis and data center vendors to provide end to end solutions through strategic acquisitions to fill gaps in its product portfolio. Analysis of the leading vendors have released a lightweight version of enterprise analytics platform to address a variety of edge use cases. OEM supplier of data center edge computing as the opportunity to ease its revenue decline, focused on providing customized hardware and system management software to meet the needs of the Internet of Things industry. Large cloud providers will edge computing products as an extension of a cloud-based platform for things they were trying to create an ecosystem of hardware vendors to ensure that their edge calculation software certified or validated, these can run on hardware products . The industry has also seen some new suppliers that provide the cloud of the Internet of Things platform integration software platform based on the edge of computing. The following are the future of computing will shape the edges of the key market trends: the edge of AI reasoning skills will facilitate localized insight and real-time response, although data from different sources of industrial systems can be centrally aggregation and analysis using AI platform, but the deployment of machine learning reasoning on the edge model analysis and binding edge significantly improve the quality margin data. By using the data available locally, you can achieve better real-time decisions. Implementation of these inference models, especially for handling text, video and voice data streams intensive computing model will require a relatively complex edge data processing architecture. The new form will satisfy a wide range of hardware edge computing needs due to the fast processor and battery technology innovation, the market will witness the introduction of new edge devices, these devices can run data analysis, and run AI reasoning model directly on the device. It is expected in the next few years, hardware forProviding vendors will focus on the edge is calculated based on the GPU, the visual processing unit, field programmable gate arrays, application-specific integrated circuit and the processor hardware. They are designed to perform complex, computationally intensive functions hardware at the edges. In addition CONTROL ENGINEERING China Copyright , the market will focus more on improving the efficiency of electricity consumption, particularly in the gateway and embedded devices such as limited systems. 5G will accelerate distributed computing, but only for some of the introduction of the use cases 5G cellular technology may affect the I & O architects to re-evaluate their edge computing architecture, in particular the use cases for mobile properties, such as autonomous vehicles, fleet management, transport and logistics, and other things with the integrated embodiment. 5G cellular technology can achieve gigabit download and upload speeds. Therefore, some companies may send all the data to evaluate the possibility of generation of cloud or data center near real-time analysis. On the other hand, 5G may also accelerate truly distributed computing architecture deployment. However, CONTROL ENGINEERING China Copyright , the early stages of 5G cost may be too high for large-scale projects of things. Also CONTROL ENGINEERING China Copyright , 5G may not be used for non-edge connector 5G remote areas use cases. The sensor fusion improve data quality at the edges, but the edges of computing architecture would complicate the sensor fusion commonly used in personal devices, such as smart phones and personal medical device, wherein from the GPS, gyro and accelerometer data in a device comprising combination and analysis dedicated microcontroller unit, to provide users with a specific, personalized views. Sensor Fusion is also applicable to self-driving car , the data obtained from a laser radar, a camera, odometer, radar and other sensors and streamlining fusion algorithms such as Kalman filter, to provide a comprehensive understanding of the surrounding environment. Sensor fusion concept is expected to influence the future shape of the industry edge computing architecture. The ability of various combinations of data from different types of sensors together, helps to generate in the vicinity of the source of data to improve data quality. Also CONTROL ENGICopyright NEERING China In other cases, it may also provide a complete view of the edge environment. However, CONTROL ENGINEERING China Copyright , the integration of data from different sources is complicated, since it requires uptake and processing of a plurality of standardized data streams. This will force the infrastructure architects and data architects to redesign their edge infrastructure. Data analysis and AI capabilities will move up into the industrial control network is now in the Internet of Things project Control Engineering Copyright , most related to IoT analysis process is outside the control of industrial network via a gateway implementation, or through industrial server platform in the OT, or combination of both. This is because there is no industrial automation suppliers and advanced data analysis with AI-centric capabilities into industrial control systems such as PLCs and DCSs. More importantly, so far, they have not been considered a must-have capability. However, CONTROL ENGINEERING China Copyright , Gartner forecast With the IT / OT integration and the transition to industrial 4.0 Control Engineering Copyright , this situation will change gradually. Gartner expects the supplier of industrial automation and data analysis will be more in-depth cooperation between suppliers to create intelligent industrial control systems. This will provide companies a better asset control, asset life cycle management and new sources of revenue. Container will be considered the cornerstone of large-scale field deployment project things in the local gateway and edge server application may need to update or add new use cases workflow. These systems typically operate in a bandwidth-limited network. Modularity, portability, low cost, and isolation and other characteristics of the container industry to become a vertical cross-site Internet of Things project attractive option. Container ensures minimal change management overhead update application components, thus reducing overall bandwidth requirements. They also provide help to isolate a specific process environment.