Microservice architecture usage

Cloud Container Engine (CCE) is used to split the service modules into smaller service modules. The modules are independently deployed and upgraded for the convenience of application management.

Each container can use image services of different environments. When the container is started, an independent micro service host node (independent network IP) is generated. The service discovery between the upper layer service and the lower layer service can be solved flexibly through various ways such as environment variable injection, configuration file mounting, etc. In addition, various cloud services and customized micro services provided by the cloud platform can be directly integrated to form a powerful service cluster.

More importantly, the migration and replication of cluster environment with so many services is also very easy. Just select the corresponding Docker service image of each service and configure the access address between each other, a exactly the same new cluster can be built up.

Output of AI technology

It supports the good operation of deep learning on the cloud container. The container engine can provide complete, stable and highly available K8S cluster service and large-scale computing storage capacity, while grafting the GPU and FPGA capabilities. It is convenient for differential output of AI capabilities such as Baidu driverless car, and also solves the scenes of rapid deployment and startup, resource sharing and isolation, automatic task scheduling, fault recovery, standardized model training to online workflow for Baidu AI capability landing in enterprise users.

Improve resource utilization and reduce cost

In some cases, users only need to perform one-time tasks, such as calculating a certain result. If the traditional service mode is adopted, the server needs to run all the time, resulting in a huge waste of resources. CCE can be used to create and destroy containers quickly, which can meet the demand for payment scenario. As long as the Docker image of the one-time program is made, when it shall be run, the container is created in real time to perform tasks, and the container automatically exits and releases resources after the program execution.

Continuous integration/continuous delivery

The excellent DevOps environment provided by continuous integration and continuous delivery greatly improves the efficiency of software release.

Continuous integration: Continuous integration emphasizes that developers should build and (unit) test new code as soon as it is submitted. Based on the test results, we can determine whether the new code and the original code can be correctly integrated together.

Continuous delivery: Continuous delivery deploys the integrated code to the running environment on the basis of continuous integration.

By deploying services on Cloud Container Engine service, developers can immediately build, test, package and integrate new codes after they submit them on git or other code platforms. On the basis of continuous integration, the integrated code is deployed to the pre-release environment and the existing network environment.