Scribd
Upload
35 views

Why Containerd in Kubernetes

Kubernetes has deprecated docker as a container runtime since version V.120 due to docker leading to more latency than other options like cri-o and rkt. Using containerd directly is faster than going through the docker daemon and CLI when creating containers in kubernetes.

Uploaded by

jyoti.mgph
Copyright
© © All Rights Reserved
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
35 views

Why Containerd in Kubernetes

Kubernetes has deprecated docker as a container runtime since version V.120 due to docker leading to more latency than other options like cri-o and rkt. Using containerd directly is faster than going through the docker daemon and CLI when creating containers in kubernetes.

Uploaded by

jyoti.mgph
Copyright
© © All Rights Reserved
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 3

Why are we using containerd over docker in kubernetes?

Kubernetes has deprecated docker as a container runtime since version V.120. This is
due to the reason that using docker leads to more latency compared to any other. There
are a few other options such as cri-o, rkt container engine.

Here, kubelet acts as our client which sends requests if we want to deploy
pods/containers. CRI is a layer that acts between container runtime and kubelet.
Container runtime is what we know as docker, containerd, cri-o, rkt, etc.
Between kubelet and CRI a grcp protocol is formed. Once, CRI sends this request to
container runtime, it creates the necessary containers and informs kubelet.

Whereas if we look at the container creation workflow while using docker, it looks like
this:

VISUALPATH: #205, 2nd.Floor, Nilgiri Block, Aditya Enclave,

Ameerpet,Hyderabad
Here, docker also uses containerd in the backend to create a container. Since if we are
using docker, it first has to go through the docker CLI and then to the docker daemon
and finally to containerd, kubernetes has decided to use containerd by default as it is
much faster than using a middleman in docker.

Note: We still have a great use for docker as creating custom images and adding them
to the docker registry is much easier and also using a docker engine in our local system
is easier than creating a whole cluster like kubernetes.
In containerd, there has been a few changes from v1.0 to v2.0:

VISUALPATH: #205, 2nd.Floor, Nilgiri Block, Aditya Enclave,

Ameerpet,Hyderabad
The intermediate stem in V1.0 where it is passed through the CRI layer has been
modified as CRI is now available as a plugin in containerd from v2.0.

VISUALPATH: #205, 2nd.Floor, Nilgiri Block, Aditya Enclave,

Ameerpet,Hyderabad

You might also like