Edge computing nodes are built based on CDN edge nodes, covering most parts of the country, with full coverage of the three major operators. Sink computing power to each city-level node to provide computing power resources closer to users.
Then some friends may ask, CDN also provides services for users nearby, what is the difference between edge computing nodes and CDN?
1. What is a CDN
In fact, CDN is a content distribution network. It mainly solves the problem of slow network access due to long distances. Therefore, CDN nodes are usually deployed near the user, so that when users access content resources, they do not need to access the central server, but directly access the nearest CDN node. Not only saves the time cost of traveling long distances, but also achieves network acceleration.
But CDN also has its own shortcomings. Take the live broadcast scene we are familiar with as an example: in the live broadcast scene, the live broadcast content is generally uploaded to the center for transcoding, and then delivered to the audience through CDN acceleration. Here, the main role of CDN is to shorten the transmission distance between the central server and users.
CDN is mainly used for caching and accelerating content, and it has no computing power, so the data must be transmitted to the central server for computing and processing. The video transmission link is long, the cost is still high, and the data processing efficiency is still very low.
2. The relationship between CDN and edge computing
For this reason, edge cloud technology, which combines the advantages of cloud computing and CDN, was born. Edge computing is at the edge node/computer room close to the user side, and the cloud computing capability of the central cloud is lowered to the user side. Take our live broadcast scene just now as an example. If edge computing is used in the live broadcast scene, how will the cost problem involved in CDN be solved?
First of all, after the edge cloud node is used as the source station of video live broadcast, the live broadcast content can be uploaded, transcoded and streamed nearby, and can also be distributed to local users, which can well realize the quality assurance of localized live broadcast links, and the average delay It can be shortened from 40ms to 20ms, which not only ensures user experience, but also saves bandwidth costs.
Moreover, for CDN, edge computing not only adds storage and computing functions, but also a series of basic capabilities such as network and security.
3. What are the edge computing nodes?
The edge computing node BEC supports users to create resources such as edge containers, edge cloud servers, load balancing, and edge storage. Among them, edge storage includes edge object storage and edge cloud disk.
Today, I will introduce edge containers and edge cloud servers to you first.
3.1 Edge Containers
Users can create edge container applications in the BEC console. Supports customizing the container application name, selecting nodes, configuring storage volumes, configuring container mirroring, etc., and the edge container can be used after the configuration is successful.
- Edge application
Users manage computing resources by creating edge applications. The smallest container group is used as the deployment and management unit to run and manage cross-machine container groups in a cluster. An edge application can contain multiple deployments.
- deploy
Deploy the StatefulSet or Deployment corresponding to Kubernetes. If an edge service contains multiple edge nodes (such as North China Mobile and East China Unicom), then the edge service will contain multiple deployments, and each edge node corresponds to a deployment. Contains multiple container groups.
- container group
The container group corresponds to the Pod concept of Kubernetes, and the container group contains multiple containers. The container group has independent storage, IP, port, namespaces, and configuration, and is the computing power management unit of the edge computing node service.
- mirror image
Docker images are used to deploy container services. Images allow users to upload private images, or use official or Baidu smart cloud images.
- mirror warehouse
The mirror warehouse is a user-owned Docker image storage space, which runs on Baidu Smart Cloud Container Service, and Baidu Smart Cloud is responsible for ensuring reliability, security and performance.
The structural relationship of edge applications, deployments, container groups, and containers is shown in the following figure:
3.2 Edge cloud server
Through the edge cloud server, we can understand important information such as instance groups, instances, purchased resources and purchased nodes, instance group monitoring, instance monitoring, and load balancing monitoring
- instance group
An edge cloud server instance group is a resource group composed of instances. Users can organize instances into an instance group during the stage of creating a cloud server for flexible batch management. Users can add new instances to the instance group during the instance group expansion stage.
- example
The edge cloud server instance is a virtual computing environment, including the most basic computing components such as CPU and memory, and is the actual operating entity presented by the edge cloud server to each user. The BEC instance is the core concept of the edge cloud server and supports functions such as IP binding, storage, and mirroring.
- load balancing
The load balancing function can balance application traffic among multiple edge cloud servers, cope with massive access requests, and realize business horizontal expansion. In addition, load balancing can avoid single point of failure and improve business availability.
Welcome to learn more about edge computing nodes: edge computing node BEC