All Products
Search

This Product

    Container Service for Kubernetes:Comparison among Docker, containerd, and Sandboxed-Container

    Document Center

    Container Service for Kubernetes:Comparison among Docker, containerd, and Sandboxed-Container

    Last Updated:Nov 14, 2024

    Containers and images have become the industry standards for software packaging and delivery. Kubernetes has become a standard platform for building, developing, and managing containerized cloud-native applications. An increasing number of enterprises and customers deploy their applications in Container Service for Kubernetes (ACK). ACK supports the following container runtimes: containerd, Docker, and Sandboxed-Container. This topic compares these runtimes in terms of implementations, limits, and deployment architectures. This topic also compares the commonly used commands provided by Docker Engine and containerd. This helps you select a container runtime based on your business requirements and scenarios.

    Comparison in terms of implementations and limits

    Feature

    containerd

    Docker

    Sandboxed-Container

    Description

    Cluster types

    • ACK managed clusters

    • ACK dedicated clusters

    • ACK Edge clusters

    All types of of ACK clusters

    • ACK managed clusters

    • ACK dedicated clusters

    None

    Kubernetes version

    Kubernetes 1.31 and earlier

    Kubernetes 1.22 and earlier

    Note

    We recommend that you upgrade Docker to containerd for your clusters. For more information, see Change the container runtime from Docker to containerd.

    Kubernetes 1.31 and earlier

    None

    Node type

    • Elastic Compute Service (ECS) instance

    • ECS Bare Metal instance

    • Self-managed nodes (ACK Edge clusters)

    • ECS instance

    • ECS Bare Metal instance

    ECS Bare Metal

    None

    Node operating system

    • CentOS

    • Alibaba Cloud Linux

    • Windows for ACK 1.20.4 and later

    • Ubuntu and CentOS (ACK Edge clusters)

    • CentOS

    • Alibaba Cloud Linux

    Alibaba Cloud Linux Customized Edition

    • You cannot deploy both Docker and Sandboxed-Container on one node.

    • To deploy both Docker and Sandboxed-Container in the same cluster, you can create node pools that use different types of container runtimes.

    Container engine

    containerd

    Docker Engine

    containerd

    None

    Monitoring

    Supported

    Supported

    Supported

    None

    Log collection

    Supported

    Supported

    Manual configuration of sidecar containers is required to support log collection.

    For more information about the sidecar configuration, see Use CRDs to collect container text logs in Sidecar mode.

    Stdout collection

    Supported

    Supported

    Supported

    None

    RuntimeClass

    Not supported

    Not supported

    Supported (runV)

    None

    Pod scheduling

    No configuration is required.

    No configuration is required.

    You must configure pods based on the following rules:

    • For Kubernetes 1.14.x, you must add the following configuration to the nodeSelector parameter: 

      alibabacloud.com/sandboxed-container: Sandboxed-Container.runv

    • For Kubernetes 1.16.x and later, no configuration is required.

    None

    HostNetwork

    Supported

    Supported

    Not supported

    None

    exec/logs

    Supported

    Supported

    Supported

    None

    Node data disk

    Optional

    Optional

    Required (at least 200 GiB)

    None

    Network plug-in

    • Flannel

    • Terway (excluding ACK Edge clusters)

    • Flannel

    • Terway

    • Flannel

    • terway-eniip: Only terway-eniip that does not run in IPVLAN or Assign One ENI to Each Pod mode is supported.

    None

    Kube-proxy mode

    • iptables

    • IPVS

    • iptables

    • IPVS

    • iptables

    • IPVS

    None

    Volume plug-in

    CSI (excluding ACK Edge clusters)

    CSI

    CSI

    None

    Container root file system

    OverlayFS

    OverlayFS

    OverlayFS with support for disk quotas

    None

    Comparison in terms of deployment architectures

    Runtime

    Deployment architecture

    Docker

    kubelet -> dockerd -> containerd -> containerd-shim -> runC containers

    containerd

    kubelet -> containerd -> containerd-shim -> runC containers

    Sandboxed-Container V2

    kubelet->(CRI)containerd
\-> containerd-shim -> runC containers
\-> containerd-shim-rund-v2 -> runV sandboxed containers

    Comparison of the commonly used commands provided by Docker Engine and containerd

    Docker uses Docker Engine for container lifecycle management. Sandboxed-Container uses containerd for container lifecycle management. These tools provide different commands that can be used to manage images and containers. The following table describes the commonly used commands provided by Docker Engine and containerd.

    Operation

    Docker

    containerd

    docker

    crictl (recommended)

    ctr

    Query containers

    docker ps

    crictl ps

    ctr -n k8s.io c ls

    Query container details

    docker inspect

    crictl inspect

    ctr -n k8s.io c info

    Query container logs

    docker logs

    crictl logs

    None

    Run commands in containers

    docker exec

    crictl exec

    None

    Attach local stdin, stdout, and stderr to containers

    docker attach

    crictl attach

    None

    Query resource usage statistics

    docker stats

    crictl stats

    None

    Create containers

    docker create

    crictl create

    ctr -n k8s.io c create

    Start containers

    docker start

    crictl start

    ctr -n k8s.io run

    Stop containers

    docker stop

    crictl stop

    None

    Delete containers

    docker rm

    crictl rm

    ctr -n k8s.io c del

    Query images

    docker images

    crictl images

    ctr -n k8s.io i ls

    Query image details

    docker inspect

    crictl inspecti

    None

    Pull images

    docker pull

    crictl pull

    ctr -n k8s.io i pull

    Push images

    docker push

    None

    ctr -n k8s.io i push

    Delete images

    docker rmi

    crictl rmi

    ctr -n k8s.io i rm

    Query pods

    None

    crictl pods

    None

    Query pod details

    None

    crictl inspectp

    None

    Start pods

    None

    crictl runp

    None

    Stop pods

    None

    crictl stopp

    None