In a 2019 Forrester survey found that 53 percent of respondents expect that in the next three years, they will analyze complex data sets at the margins.
The compound annual growth rate to calculate the size of the relevant market from 2018 to 2022 will exceed 30% of the world's edge, to 2022, the market size will exceed trillion edge computing, and cloud computing market size comparable. Edge computing has become an important trend of things.
Networking and computing edge
Things (English: Internet of Things, acronym IoT) is a common definition refers to the Internet, telecommunications networks and other traditional information carrier, so that all can exercise independent function of common objects to achieve interoperability of networks.
In the Internet of Things, everyone can use electronic tags to real objects Internet link, the Internet of Things can find out their specific locations. Can be networked with a central computer through the objective for machinery, equipment, personnel centrally manage, control, remote control can also be carried out, and the search for the location of household equipment, cars, and other items were stolen to prevent similar automated control systems, and through the collection of these little things data can be gathered into a big final data, include redesigned to reduce road traffic accident, urban renewal, disaster prediction and control of epidemics and so on major social change, and to achieve material objects associated. So, in fact things are important means to realize the digital transformation of the industry, and also spawned a new industry ecosystem and business models.
So what is it calculated edge? Calculating the edge layer is disposed directly treated by adding between the terminal edge of the network layer and the data center, then the data without the cloud, the cloud to substantially alleviate the pressure. On this basis, we can provide the nearest edge of the smart services to meet the industry agile digital connection, real-time business, data optimization, application intelligence, security and privacy protection and other critical needs. Farewell conventional finishing to the central server, the post-processing mode, the edge calculation processing may be implemented for data on each device within the network directly.
Means of calculating an edge of things can enhance intelligence, prompting things root in various industry vertical.
What are the things intelligent edge computing
Things intelligent edge computing (IoT Edge) is a set of deployment and operation of intelligent systems on the edge node devices, cloud computing will expand to the edge nodes, intelligent calculations are performed in the field, it is to achieve a distributed autonomy, an important support for automation control . The user can obtain, for example, computing node device access management, data processing nearby, and the like AI inference capabilities deployed in the field by the edges of the device.
Intelligent Edge calculated things, the customer can easily access edge devices, work quickly cloud data on conventional equipment without networking capabilities and mass sensors. Computing applications through a variety of edge, customers can conduct real-time data processing at the job site, real-time decision-making, real-time response, real-time control and local data storage security.
Platform, users can edge devices based on the actual architecture and the operating system installed, the platform completed registration, and by mounting the local installation package, to complete the authentication by the network edge node to the intelligent edge jingdong cloud AI was calculated.
- Reliable sub-device management : platform can help users to easily add child devices, configure and update device control sub-devices.
- A key deployment edge applications : providing an apparatus for connecting devices and application services for intelligent computing edge applications, users can make use of a platform key application deployed on the edge node by deploying different applications can optimize and improve edge device ability.
- Personalized application configuration : Things smart edge computing platform provides message routing, equipment configuration data can be easily filtered, and different applications for different data sources specified device can switch at any time. At the same time, users can also own AI model deployment to the edge nodes for intelligent inference and analysis.
Jingdong cloud and AI Things intelligent edge computing services
Practice: vision-based object detection
A ready condition
1. Go Things engine console, create a product template using IP cameras and register a camera device in the product.
2. Register your edge node. Once registered, you will install the system package installed on your edge node devices, networking certification complete activation.
Click on the left edge of the application, the deployment of IP camera equipment services (to support IP cameras connected) to the upper edge of the current node, the current IP cameras support ONVIF service agreement.
Second, access to the IP camera apparatus and an object detection application install
点击左侧边缘应用，部署对象检测应用至当前边缘节点上。
进入当前边缘节点详情页面，点击子设备管理->添加子设备。将该摄像头添加至边缘节点上，IP摄像头和边缘节点保持在同一局域网内。
点击消息路由->新建消息路由，创建一个消息路由，将刚刚添加的摄像头作为消息源，对象检测应用作为目的端。
三、开启摄像头并运行对象检测应用
1.点击子设备管理标签页，更新摄像头设备影子。
streaming 设置为 on （打开摄像头），其他可保持初始值。
2.点击对象检测应用的更新影子。按实际需求，填写相关参数值后，更新影子。
点击边缘应用标签页，启动对象检测应用。
四、查看对象检测结果
您可以登录控制台。进入到创建的用于存储边缘计算结果的OSS Bucket目录下查看当前应用的检测结果。
结果文件包括：
- 图片文件：检测到目标的对应图像文件pic_20191105140035.jpg
- 文本文件：检测结果信息，包括检测到的物体label，置信度以及在图片中的坐标位置，例如： pic_20191105140035.txt。文件格式如下：
{"lable":"mouse","score":0.56487995,"left":227,"right":347,"top":289,"bottom":411} - 视频文件：如在对象检测应用影子中设置了开启录像，则在检测到目标后，会持续录制一段时间（时间请配置应用影子参数duration。默认为10秒）的视频。例如：video_20191105140045.mp4。
业务场景
目前智能边缘计算的对象检测应用可适用于如下场景：
安防监控——动作识别，人体检测，侵入检测，危险品检测；
Retail Intelligence - intelligent containers of goods recognition, face pay, thermal analysis;
Industrial inspection - defect detection, the detection material;
Picking - package identification, a single face identification, product identification;
Intelligent Transportation - license plate recognition, traffic detection, behavior recognition.
Finally draw the focus, and AI Jingdong cloud computing Things smart edge in the current Free! Free in! Free in!
Welcome to click [ link ] use.
While welcoming click on the " Jingdong cloud " for more exciting content.
