Kubernetes vs Docker – check the differences

Dear friends, today we will discuss about Kubernetes vs Docker and we will check what is basic difference between kubernetes and docker. So, let’s start and see these differences.

We should know about kubernetes and docker before checking difference.

What is Kubernetes?

Kubernetes is an open-source container management software developed in the Google platform. It helps you to manage a containerized application in various types of physical, virtual, and cloud environments.

What is Docker?

Docker is an open-source lightweight containerization technology. It allows you to automate the deployment of applications in lightweight and portable containers.

Kubernetes vs Docker

kubernetes vs docker

How Docker swarm varies with Docker?

  1. Docker Swarm (Classic)
  2. Docker Swarm Mode

Docker Swarm mode is essentially diverse from Classic Swarm.

The basic differences are listed below: –

Docker Swarm ModeDocker Classic Swarm
Docker Swarm Mode comes included hooked on Docker EngineDocker Swarm is a GITHUB repository and comes as a disconnect project. It is NOT included hooked on Docker Engine
Comes through inherent Service DiscoveryRequire external KV store lied on Consul & etc.
Comes with inbuilt feature like: –
Ranging
Systematic Updates
Facility Discovery
Load-Comparison
Routing Network
Topological Situation  
Need of inherent feature like Load Comparison, Scalability/Ranging, Routing Network etc
Secured Control & Information PlaneControl Plane and Data Plane are anxious

Difference between Kubernetes and Docker

Set up and installation

Kubernetes

  1. It requires a progression of manual strides to arrangement Kubernetes Master and specialist Hubs segments in a bunch of nodes.
  2. Kubernetes can run on different stages: from your PC to VMs on a cloud supplier, to a rack of uncovered metal servers. For setting up a solitary hub K8s bunch, one can utilize Minikube.
  3. Kubernetes support for Windows server is under the beta stage.
  4. Kubernetes Client and Server bundles should be redesigned physically on all the frameworks

Docker

  1. Installing Docker involves joke order on Linux Platform like Debian, Ubuntu, and CentOS.
  2. To install a solitary node Docker Swarm or Kubernetes group, one can send Docker for Mac and Docker for Windows.
  3. Docker keeps official help for Windows 10 and Windows Server 2016 and 1709 It’s so natural to overhaul Docker Engine under Docker for Mac and Windows through only 1 snap.

Working in two systems

Kubernetes

  1. Kubernetes works at the application level instead of at the equipment level. Kubernetes intends to help a very assorted assortment of outstanding tasks at hand, including stateless, stateful, and information preparing remaining burdens. On the off chance that an application can run in a compartment, it should run extraordinary on Kubernetes.
  2. Kubernetes can run on Docker yet expects you to know the order line interface (CLI) particulars for both to get to your information over the API.
  3. There is a Kubernetes customer called kubectl which converses with Kube API which is running on your lord hub.
  4. Not at all like Master parts that generally run on a solitary hub (except if High Availability Setup is unequivocally expressed), Node segments run on each hub.
  5. kubelet: specialist running on the hub to investigate the holder wellbeing and report to the ace just as tuning in to new orders from the Kube-API server
  6. Kube-intermediary: maintains the system rules
  7. container runtime: software for working on the containers (e.g. Docker, rkt, runc.)

Docker

  1. Docker Platform is accessible as two versions: – Docker Community Edition Docker Enterprise Edition
  2. Docker Community accompanies network-based help discussions while Docker Enterprise Edition is offered as big business-class support with characterized SLAs and private help channels.
  3. Docker Community and Enterprise Edition both drop naturally with Docker Swarm Mode. Moreover, Kubernetes is bolstered under Docker Enterprise Edition.
  4. For Docker Swarm Mode, one can utilize the Docker Compose document and use Docker Stack Deploy CLI to send an application over the bunch hubs.
  5. The ‘docker stack’ CLI conveys another stack or update a current stack. The customer and daemon API must both be at any rate 1.25 to utilize this order. One can utilize the docker adaptation order on the customer to check your customer and daemon API variants.

For move about kubernetes vs docker you can Click-Here

Logging and Monitoring

Kubernetes

  1. Kubernetes gives no local stockpiling answer for log information, yet you can coordinate many existing logging arrangements into your Kubernetes bunch. Not many of well-known logging apparatuses are recorded underneath:
  2. Fluentd is an open-source information authority for a brought together logging layer. It’s written in Ruby with a module arranged design.
  3. It assists with gathering, course, and stores various logs from various sources. While Fluentd is enhanced to be effortlessly expanded utilizing module engineering, a familiar piece is intended for execution.
  4. It’s minimized and written in C so it very well may be empowered to moderate IOT gadgets and stay sufficiently quick to move a colossal amount of logs. Besides, it has worked in Kubernete’s support. It’s a particularly reduced apparatus intended to ship logs from all hubs. Different apparatuses like Stackdriver logging gave by GCP, Logz.io and other outsider drivers are accessible as well.

Monitoring

  1. There are different open-source apparatuses accessible for Kubernetes application checking like:
  2. Heapster: Installed as a unit within Kubernetes, it accumulates information and occasions from the compartments and cases inside the bunch.
  3. Prometheus: Open-source Cloud Native Computing Foundation (CNCF) venture that offers ground-breaking questioning abilities, representation, and caution.
  4. Grafana: Used related to Heapster for imagining information inside your Kubernetes condition.
  5. InfluxDB: An exceptionally accessible database stage that stores the information caught by all the Heapster units. CAdvisor: centers around holder level execution and asset use. This comes installed legitimately into kubelet and should consequently find dynamic holders.

Docker

  1. Logging driver modules are accessible in Docker 17.05 and higher. Logging abilities accessible in Docker are uncovered as drivers, which is helpful since one gets the chance to pick how and where log messages ought to be delivered
  2. Docker incorporates numerous logging components to assist you with getting data from running holders and administrations. These systems are called logging drivers.
  3. Every Docker daemon has a default logging driver, which every holder utilizes except if you design it to utilize an alternate logging driver.
  4. Notwithstanding utilizing the logging drivers included with Docker, you can likewise execute and utilize logging driver modules.
  5. To design the Docker daemon to default to a particular logging driver, set the estimation of log-driver to the name of the logging driver in the daemon.json record, which is situated in/and so forth/docker/on Linux has.
  6. The accompanying model expressly sets the default logging driver to Syslog:

{

“log-driver”: “Syslog”

}

At the point when you start a compartment/container, you can design it to utilize an alternate logging driver than the Docker daemon default, utilizing the – log-driver banner. If the logging driver has configurable choices, you can set them utilizing at least one occurrence of the – log-pick <NAME>=<VALUE> banner. Regardless of whether the holder utilizes the default logging driver, it can utilize diverse configurable alternatives.

Size

Kubernetes

As per official page of Kubernetes documentation K8s v1.12 buck up clusters with up to 5000 hubs based on the underneath condition

  1. Not more than 5000 nodes
  2. Not more than 150000 total pods
  3. Not more than 300000 total containers
  4. Not more than 100 pods per node.

Docker

Agreeing to the Docker’s web journal post on scaling Swarm clusters has been distributed amid Nov 2015 and Docker Swarm has been scaled and execution tried up to 30,000 containers and 1,000.

Specs

  1. Innovation backend: diplomat
  2. 1,000 nodes
  3. 30 containers for each node
  4. Manager: AWS m4.xlarge (4 CPUs, 16GB RAM)
  5. Hubs: AWS t2.micro (1 CPU, 1 GB RAM)
  6. Container icon: Ubuntu 14.04

Result

Percentile  API Reaction Time Forecast Postpone

50th     150ms              230ms
90th      200ms             250ms
99th      360ms             400ms

That’s all we have seen this session about Kubernetes vs Docker

If you want to install kubernetes cluster on CentOS 7 VM you can click on below link

How to install kubernetes cluster on CentOS 7 VM

Leave a Reply