Lecture Review丨New Possibilities of Cloud-Edge-Device Collaboration Based on OpenYurt and EdgeX

Author: EdgeX Foundry

The 2022 EdgeX China Challenge and the EdgeX Special Competition of Zhongguancun International Frontier Technology Innovation Competition officially kicked off. This competition is divided into two tracks: medical, education, consumer industry track and energy, industry, supply chain track. The competition is committed to building a learning and sharing platform for the Internet of Things and edge computing. Based on open source technologies such as EdgeX Foundry and OpenYurt, it aims at multiple application scenarios in different tracks, and uses shared technology investment to solve industry technical problems.

In order to help contestants better understand and use related technologies, this competition will continue to carry out 3 rounds of technical training from July to September, covering different levels of elementary, middle and advanced, to help developers systematically learn the knowledge of intelligent edge systems. We invited technical experts from Intel, VMware, Alibaba Cloud and other institutions to share.

Edge Computing and Edge Cloud Native

First, Xiong Feng explained the detailed definitions of edge computing and edge cloud native for us.

Edge Computing:

• It is a computing method that deploys workloads at the edge ;
• It is a decentralized computing architecture that performs application computing and data storage close to things, data sources or users;
• Core goal: rapid decision-making ;
• Four drivers: latency/determinism, data/bandwidth, limited autonomy, privacy/security ;
• At present, it is mainly used in automobile, agriculture, transportation, medical care and other industries.

Legend: IoT and edge computing

Edge computing has two definitions of tiered architecture based on sensitivity to latency: the Gartner definition and the IDC definition.

Legend: Edge computing layered architecture defined by Gartner

In the Gartner definition, edge computing is stratified by latency sensitivity as:

• Near Edge: Non-standard server or device, in the place closest to the end side;
• Far Edge: Standard IDC, which can be divided into three types: IDC, MEC, CDN, etc.;
• Cloud: Public cloud or proprietary cloud service, characterized by centralized resource management and centralized management.

Legend: Edge computing layered architecture defined by IDC

In the IDC definition, edge computing is stratified by latency sensitivity as:

• Heavy Edge: data center dimension; centralized computing platform (CDN, self-built IDC);
• Light Edge: A low-power computing platform, suitable for IoT scenarios such as industrial control, data processing, and transmission.

Legend: Market size of public and non-public cloud services from 2015 to 2024 (forecast)

At present, cloud native uses an open and standard technology system to quickly build and run a highly elastic, fault-tolerant, and easy-to-manage system, helping enterprises maximize the ability to utilize the cloud and maximize the value of the cloud. Cloud native restructures the way enterprises go to the cloud, and it has become the norm for enterprises to go to the cloud.

Legend: Edge cloud native architecture integrating cloud, edge and end

Cloud native accelerates multi-cloud, cloud-edge integration, and builds an edge cloud-native architecture that integrates cloud, edge, and device. Among them, edge computing cloud-edge integrated infrastructure is widely used because of the following characteristics:

• Provide the same functionality and experience on the edge infrastructure as on the cloud;
• Cloud-side operation and maintenance collaboration, computing power mixing, network collaboration, etc.;
• DevOps collaboration on the cloud side, rapid business edge expansion;
• Device twinning enables cloud-edge-device business integration.

OpenYurt cloud native edge computing platform architecture

Expert Xiong Feng also introduced the OpenYurt cloud-native edge computing platform architecture.

OpenYurt is a CNCF sandbox project that provides cloud-edge collaborative computing capabilities for cloud management and edge autonomy . By exploring the standardization of cloud native biological models with the community, it can expand to support IoT frameworks such as EdgeX and LinkEdge , support multiple IoT protocols, and achieve device twinning capabilities .

Legend: OpenYurt cloud-native edge computing platform architecture

OpenYurt is a K8s-based edge computing cloud-native intelligent platform project, 100% compatible with K8s API, open sourced in May 2020, and entered into the CNCF sandbox in September of the same year.

OpenYurt Philosophy: Extending your native Kubernetes to edge

At present, OpenYurt has cooperated with VMware, Intel, Sangfor, China Merchants Group, Zhejiang University, Tianyi Cloud and other edge computing industry-university-research institutions to jointly promote the development of OpenYurt and strive to become the de facto standard for edge computing cloud-native. As a cloud-native PaaS kernel for edge computing, OpenYurt has covered dozens of industries and served millions of CPU cores.

OpenYurt has complete cloud-side collaboration functions, which are divided into cloud-side operation and maintenance collaboration and cloud-side data collaboration.

Legend: OpenYurt cloud-side operation and maintenance collaboration

Legend: OpenYurt cloud-edge data collaboration

OpenYurt can also work with other management methods for cluster management.

Legend: OpenYurt DevOps Collaboration

Legend: OpenYurt unit management

Legend: OpenYurt unit management: NodePool

Legend: OpenYurt unit management: UnitedDeployment

In edge computing scenarios, cloud-edge interaction may have weak network connections. In a disconnected or weak network state, the native Kubernetes cannot restore edge services when the edge node restarts.

In order to ensure the continuity of edge services and the continuity of cross-node communication of edge services when the cloud edge is disconnected from the network, the following OpenYurt edge autonomy solutions are proposed:

• YurtHub caches node data. When the cloud edge is disconnected, all system components obtain data from YurtHub;
• When the business container is restarted, the Pod IP remains unchanged;
• The MAC address of the flannel vtep remains the same when the node restarts.

Legend: OpenYurt edge autonomy solution

Legend: OpenYurt Node Pool Governance Scheme

Cloud native biological model and edge device management practice based on OpenYurt+EdgeX Foundry

Finally, Xiong Feng introduced how to use OpenYurt and EdgeX IoT framework to realize cloud native biological model and edge device management.

Legend: OpenYurt+EdgeX, opening up the last mile of cloud-edge-device integration collaboration (EdgeX 2.1 LTS support)

OpenYurt+EdgeX combined use scheme:

• Orchestrate and deploy EdgeX Foundry using OpenYurt;
• To manage devices in the real world, services related to device management need to be abstracted to provide a cloud-native IoT model;
• OpenYurt supports the management capabilities of end devices by integrating the EdgeX Foundry device management platform;
• The application management and device management paths are unified and cloud-native.

Learn More & Related Links

[1] OpenYurt Github:

https://github.com/openyurtio

[2] Overview of OpenYurt:

https://openyurt.io/en/docs/core-concepts/architecture

[3] Tunnel and operation and maintenance coordination:

https://openyurt.io/en/docs/core-concepts/yurttunnel

[4] raven works with the network:

https://openyurt.io/en/docs/core-concepts/raven

[5] Node pool and unitization:

https://openyurt.io/en/docs/core-concepts/yurt-app-manager

[6] Cloud-native device management:

https://openyurt.io/en/docs/core-concepts/yurt-device-controller**https://github.com/openyurtio/yurt-device-controller/blob/main/docs/yurt-device-controller -tutorial-v2.md

The 2022 EdgeX China Challenge has been grandly opened on August 3, and the EdgeX Chinese community will bring you more edge computing lectures and event progress updates in the coming days. In this midsummer, let us show our style together, start from the fields we have learned, provide more and better solutions for the Internet of Things and edge computing markets, and look forward to a better technological future.

Welcome to scan the QR code below to register for the competition