Plumbing the Cloud for Containers

9

Container systems are becoming easier to install, but setting up and managing production container platforms can still be challenging. In his presentation at CloudNativeCon in Seattle, Michael Friis, Product Manager at Docker, will explain what it takes to install and maintain container platforms on public clouds and describe some important considerations for choosing load balancers, logging, and storage solutions. Here, Friis gives us a preview of his talk and shares some recommended practices.

 

Linux.com: What capabilities or features are unique to Docker for AWS and/or Docker for Azure?

Michael Friis: The objective of Docker for AWS and Azure is to provide a very easy/simple way to provision and configure fully featured swarms of Docker Engine in a native experience for that target IaaS provider. Docker for AWS and Azure use the respective infrastructure installers (Cloudformation and ARM) to build simple installation experiences, and integrate with/deploy with underlying infrastructure features like load balancing and scaling in a way that feels natural and native to Azure and AWS operators.

 

Linux.com: Who should be using Docker for AWS and Docker for Azure, and why?

Michael: Anyone that would like to try out Docker 1.12 and built-in orchestration (aka swarm-mode) can get a working setup in minutes with Docker for AWS and Docker for Azure. Docker for AWS and Docker for Azure are still in beta, but they’re already great for trial and prototype Docker deployments on IaaS.

 

Linux.com: What makes setting up a production container system or PoC challenging?

Michael: There are a few items to consider when setting up a production container environment:

  • Getting security right: Networking and firewalls, configuring certificates and encryption

  • Understanding the  idiosyncrasies of your selected  underlying infrastructure, and mapping those the desired end result in the container environment.

  • Understanding how clustering software is bootstrapped, and what  ancillary software is required for the environment (like K/V-stores)

 

Linux.com: What are a few of the best practices for installing and maintaining easy-to-use container platforms on public clouds?

  • Keep track of the dependencies you take on the underlying infrastructure! Taking advantage of unique and differentiating features of a particular IaaS is great, but do so in a way that doesn’t compromise portability. Docker is a great way to help achieve this.

  • Make sure your setup is automated and scriptable. Even if you have just one production cluster, you’ll want to setup test and staging environments to try out upgrades and similar operational tasks. Make plans for upgrades: Both of the operating system (choose a minimal one with less software, and keep the complexity in container images that are much easier to upgrade), of the clustering system and of the containerized apps.

 

Registration for this event is sold out, but you can still watch the keynotes via livestream and catch the session recordings on CNCF’s YouTube channel. Sign up for the livestream now.