Edge computing refers to the use of an open platform that integrates network, computing, storage, and application core capabilities on the side close to the source of objects or data to provide the most end-to-end services nearby. Its applications are initiated on the edge side to generate faster network service responses and meet the basic needs of the industry in real-time business, application intelligence, security and privacy protection. Edge computing sits between, or on top of, physical entities and industrial connections.
According to the size of computing power, edge computing can be divided into two categories:
- Light edge : limited computing power, poor network conditions, single business, large number, geographically dispersed, such as smart communities, Internet of Vehicles, drones, etc.
- Heavy edge : The computing power is relatively sufficient, the network conditions are relatively stable, the business is complex, and there are certain requirements for reliability and security, such as 5G MEC, industrial Internet, and smart cities.
FabEdge is a container network solution based on kubernetes and focusing on edge computing scenarios. It solves the problems of complex network management in edge computing scenarios, fragmentation and non-communication, lack of topology awareness, and inability to provide nearby access. Business collaboration between cloud-edge and edge-to-edge. FabEdge supports lightweight edge nodes managed by edge computing frameworks such as KubeEdge, SuperEdge, and OpenYurt. In the newly released V0.4 version, the support for heavy edge, that is, edge cluster, is added to complete the full coverage of all edge scenarios.
As an example in the above figure, there are three clusters. The cluster blue is the host cluster, which is responsible for managing the communication of other clusters; the clusters red and green are two member clusters, which will report the network configuration information of this cluster to the host cluster blue. After adding clusters red and green to community1, FabEdge will automatically establish a tunnel between clusters red and green, allowing mutual access between pods and services between the two clusters.
The interaction process of FabEdge multi-cluster communication is shown in the following figure:
1. First create clusters green and red in the host cluster, and obtain the corresponding tokens.
2. Use the obtained token to register clusters green and red.
3. The clusters green and red report the network endpoint information of the cluster to the host cluster.
4. Add clusters green and red to a community.
5. The clusters green and red regularly pull the remote endpoint information from the host cluster.
6. According to the community information, the host cluster sends the relevant endpoint information for the clusters green and red to the member cluster operators.
7. The operators of member clusters green and red update the configmap for their connectors.
8. The connectors of member clusters green and red initiate tunnels to each other according to their own configmaps.
9. After the tunnel is successfully established, the member clusters green and red can communicate with each other.