Component Overview The Cloud Container Engine (CCE) includes various components such as networking, storage, monitoring, and logging, catering to diverse business needs. You can install, upgrade, or remove components as needed.
Cluster components managed by Cloud Container Engine (CCE) include system components and optional components:
System components: Pre-installed components available by default when creating a Kubernetes Cluster in CCE.
Optional components: Additional components that can be installed during Kubernetes Cluster creation in CCE to expand its functionality.
Component name
Types
Component introduction
Kube APIServer System components
The Kubernetes cluster gateway acts as the main API interface for interacting with Kubernetes clusters.
Kube Scheduler System components
The Kubernetes cluster scheduler manages the scheduling of cluster resources.
Kube Controller Manager System components
The Kubernetes resource manager oversees the management of internal cluster resources.
CCE Virtual Kubelet Optional components
The CCE virtual node component enables quick access to BCI elastic container capabilities.
Component name
Types
Component introduction
CCE Ingress Controller Optional components
Implement K8s Ingress rules using Baidu AI Cloud Application Load Balancer (Application BLB) to provide advanced Layer 7 network load balancing functionality.
CCE Ingress Nginx Controller Optional components
Utilize Nginx as a reverse proxy, with Baidu Load Balancer and HTTP caching, dynamically updating Nginx configuration files based on ingress resource definitions.
NodeLocal DNSCache Optional components
Enhance DNS resolution performance by deploying a DNS caching proxy as a DaemonSet on cluster nodes.
CoreDNS System components
Server for Kubernetes cluster domain name resolution.
CCE LB Controller System components
Harness Baidu AI Cloud Application Load Balancer (Application BLB) to implement Kubernetes LoadBalancer Service functionality, offering Layer 4 (TCP/UDP) traffic routing.
CCE Network Plugin System components
CCE network components for the Kubernetes cluster.
Component name
Types
Component introduction
CCE CSI BOS Plugin Optional components
Enable Baidu AI Cloud BOS storage mounting in pods using PV/PVC.
CCE CSI CDS Plugin Optional components
Enable Baidu AI Cloud’s CDS storage mounting in pods using PV/PVC and support dynamic PV creation.
CCE CSI PFS Plugin Optional components
Enable Baidu AI Cloud’s PFS storage mounting in pods using PV/PVC along with dynamic PV provisioning.
CCE CSI PFS L2 Plugin Optional components
Support mounting Baidu AI Cloud's PFS L2 storage in pods with PV/PVC and dynamic PV provisioning.
Component name
Types
Component introduction
CCE Log Operator Optional components
Collect and forward cluster log information to the Baidu Log Service (BLS) for persistent storage.
CCE Node Problem Detector Optional components
Monitor and detect various anomalies on nodes in real time, reporting results to the cluster.
CCE Node Remedier Optional components
The node auto-repair system enables automated self-healing for node failures.
CCE Onepilot Optional components
Facilitates integration of multi-language probes, allowing automatic APM probe injection and Metric reporting environment setup in K8S Cluster CCE.
Component name
Types
Component introduction
CCE Image Accelerate Optional components
Provide image acceleration for Kubernetes clusters by combining Cloud Container Registry (CCR)’s image acceleration features to support partial downloads and live decompression, greatly reducing container startup times.
CCE P2P Accelerate Optional components
Support concurrent downloading of large-scale GB-level container images in clusters, significantly boosting image pull speeds.
CCE Credential Controller Optional components
Enable password-free pulling of CCR Enterprise Edition instance images in a K8S Cluster (CCE).
Component name
Types
Component introduction
CCE Descheduler Optional components
Trigger rescheduling for pods located on high-load nodes within the cluster.
Component name
Types
Component introduction
CCE GPU Manager Optional components
Advanced scheduling features, such as GPU architecture awareness, allow sharing and isolation of GPU computing power and memory.
CCE Deep Learning Frameworks Operator Optional components
Provide distributed model training, deployment, and other features, supporting TensorFlow, PyTorch, MXNet, and PaddlePaddle frameworks.
CCE RDMA Device Plugin Optional components
Supports inter-host data transmission for containers using a high-performance RDMA network.
CCE AI Job Scheduler Optional components
Optimize for AI training scenarios with support for batch scheduling, queue management, and other functionalities.
Fluid Optional components
A Kubernetes-native, open-source distributed data orchestration and acceleration engine tailored for big data and AI applications in cloud-native environments.
CCE PaddleFlow Pipeline Optional components
A cloud-native AI workflow component streamlining AI job orchestration, templating AI training processes, and boosting experimental productivity.
CCE NPU Manager Optional components
Support sharing and isolation of Kunlun Chip NPU computing power and memory resources.
CCE Ascend Mindx DL Optional components
The deep learning component built on the Ascend AI Processor provides core functionalities, including processor scheduling and cluster performance testing.
Component name
Types
Component introduction
CCE QoS Agent Optional components
The container service quality (QoS) enhancement component delivers strong resource isolation on a single machine, boosting cluster utilization and ensuring stability.
CCE CronHPA Controller Optional components
Enable scheduled scaling of cluster workloads using customizable rules.
CCE BackUp Controller Optional components
A cluster extension component built on the open-source Velero, offering capabilities for resource backup and recovery.
