Skip to content

Latest commit

 

History

History
122 lines (67 loc) · 2.94 KB

slides-part1.md

File metadata and controls

122 lines (67 loc) · 2.94 KB
Raw

[fit] Introduction to Docker

Part 1 — What is Docker and When to use it







Badar Ahmed

^ Hello. Welcome to this introductory course on Docker.

^ I am Badar Ahmed.

^ And today I will be presenting part 1 of this course, where we will cover some background on Docker and when to use it.

What is Docker?

right

  • New technology launched in 2013.

    ​

  • Builds on top of Linux kernel features for OS level virtualization.

    ​

  • Isolated containers where processes can run without impacting each other (process, memory, filesystem isolation).

What problems does Docker solve?

  • Shipping code is hard. Docker makes it easier. ​
  • When things break in production, most common developer excuse is "Works on my laptop". ​
  • Docker provides a standardized container that works for shipping software to different environments.

^ Docker solves these problems by packaging code in a universal format that can be run across different environments in isolated containers.

^ Isolated containers allow us to be sure that a package or its dependencies are not mixing together on disk, in memory or in any other way.

When to use Docker?

  • Anywhere where you are deploying software between multiple environments. ​
  • Docker provides an easy to use and portable way to run same software anywhere. ​
  • You can run same container on your laptop, same in CI environment and same in production environment.

When to use Docker?

  • Docker containers are order of magnitude faster than VMs. ​

  • They can be ported on virtually any environment that runs relatively latest version of the Linux Kernel.

    ​

  • Can be deployed on Public Cloud (AWS/GCP/Azure), Private Cloud or on Bare Metal.

When NOT to use Docker?

  • Docker ecosystem is changing very fast as of right now.

    ​

  • If you want something mature or stable this is not it.

    ​

  • Fully leveraging Docker containers does require rethinking how to refactor or redesign code deployment pipelines.

Container Ecosystem

right

  • Docker is the defacto container engine & format.

  • Rocket by CoreOS is another alternative.

  • Lower level container primitives like cgroups & LXC are harder to use.

  • Docker provides a really nice API on top of those lower level primitives.

Container Ecosystem

right fit

  • A lot of ecosystem around containers is trying to solve problems of container scheduling, orchestration, networking & storage.

    ​

  • Major container scheduler/orchestrators are Docker Swarm, Nomad, Kubernetes & Mesos.

How does Docker affect DevOps?

  • Less sysadmin work.
  • Less variance between developer local environment and prod environment.
  • Enables more DevOps, since developers can easily package their own code!
  • Allows easier integration and end to end testing.

[fit] Thanks!