Containerization with Microsoft Azure

Editor's Notes

• #3: A virtual machine is -- well -- a virtualized machine created and managed by a hypervisor such as VirtualBox or Hyper-V. Even though a VM runs on a machine that has an operating system, each VM requires its own complete operating system, even if it's the same operating system as the host OS. VMs offer a very high degree of isolation, but at a cost: longer startup times, lower portability (ever tried to move a 127 GB virtual hard disk, or VHD, from one PC to another?), and higher memory requirements. Containers, by contrast, leverage the operating system that is already in place but offer nearly as much separation. RAM requirements are lower since the OS isn't being duplicated in each container, and cost is lower, too, because while cloud platforms typically charge for each VM, a single VM can host multiple container instances.
• #4: Docker (www.docker.com) isn't the world's only containerization platform, but it is the most popular. It is free, open-source, and Linux-based, with Windows support (Windows Server 2016) in the works. It has earned massive mindshare in the developer community. And with Azure Container Service, you can deploy Docker containers to Azure with minimal effort. Moreover, Docker containers are easily moved between Azure and Amazon Web Services (AWS), affording developers portability between cloud platforms.
• #5: Docker utilizes a client-server architecture. You execute Docker commands through a Docker client such as the Docker CLI or Kitematic. The client uses REST commands to communicate with the Docker daemon running on a Docker host such as the Azure Container services. These commands can be used to push, pull (docker pull), and create Docker images, to run them in containers, and to manage those containers. Images can be built with the docker build command, and they can be stand-alone, or they can "inherit" from other images. Images are stored in Docker registries, which can be public or private, local or remote. Docker Hub is a popular public registry that is managed by Docker; it contains a "huge collection" of images that anyone may use. The docker run command runs a container using an image as a template.
• #6: The Docker Client, also known as the Docker CLI, is the primary tool you use to manage Docker containers. You can download container images from repositories such as Docker Hub, build container images, run container instances, list container images and instances, and much more. After connecting to Azure Container Service using SSH, you can use port forwarding to execute commands locally that act on an Azure Container Service running in the cloud. In this example, the -H switch used with the docker commands forwards commands sent to port 22375 on localhost to the Azure Container Services via SSH.
• #7: This command pulls the image named "Ubuntu" from Docker Hub (or a local registry if the image is cached there) and runs it interactively in a container. "Interactively" means standard input, output, and error are connected locally so you can provide input to the container and see its output. Of course, you are not limited to the "Ubuntu" image. You can specify other images and even create images of your own with docker build. Where the container runs depends on the context. The container can run locally in a docker host (for example, a VM on Windows), or it can remotely if you connect to a remote Docker daemon (for example, one running in Azure) via SSH tunneling and use port forwarding to forward docker commands to the daemon.
• #8: These are some of the most commonly used docker commands. You can also use docker push to push an image to a registry such as Docker Hub. Also, docker ps is often accompanied by a -a switch to list all containers, including those that are no longer running, while docker rm and docker rmi are used to delete (remove) containers and images, respectively. The docker build command uses a Dockerfile (a text file containing build commands) and a "context" -- for example, a specified directory in the file system -- to build Docker images.
• #9: From the documentation: "Azure Container Service makes it simpler for you to create, configure, and manage a cluster of virtual machines that are preconfigured to run containerized applications. It uses an optimized configuration of popular open-source scheduling and orchestration tools. This enables you to use your existing skills, or draw upon a large and growing body of community expertise, to deploy and manage container-based applications on Microsoft Azure." ACS supports Linux containers and Windows containers. The latter rely on Windows Server 2016.
• #10: Orchestration in the context of containers is the deployment and management of containers across infrastructure and networks. It provides the tools and software defined infrastructure needed to deploy containers. Containers by design are intended to be deployed in large volumes with some applications using dozen to even thousands of containers. With this type of scale, automating deployment and management of containers with Orchestration Software becomes necessary.
• #11: A container cluster is intended to have redundancy for load balancing, scalability, and high availability. A cluster is composed of one or more Master Nodes which control the orchestration for scaling and delegation of tasks to the agents as well as provide monitoring. The Agent Nodes actually run the container loads.
• #12: Kubernetes is an open-source platform for automating deployment, scaling, and operations of application containers across clusters of hosts, providing container-centric infrastructure. With Kubernetes, you are able to quickly and efficiently respond to customer demand: Deploy your applications quickly and predictably. Scale your applications on the fly. Seamlessly roll out new features. Optimize use of your hardware by using only the resources you need.
• #13: Datacenter Operating System is, as the name implies, an operating system for for a datacenter rather than for a phyiscal of virtual machine. DC/OS abstracts away the underlying hardware from applications and provides much of the needed infrastructure that would otherwise come from multiple disparate services on traditional networks. This enables operations and developers alike to deploy and maintain applications in a uniform fashion with less time and resources spent monitoring the supporting network and network services for those applications.
• #14: Docker Swarm is the orchestration engine from Docker. It has tight integration with Docker, so many of the same paradigms and tools that used with Swarm seamlessly as well as some additional features like overlay networks and Docker services that can run multiple containers spread across multiple hosts. These containers can seamlessly communicate and rapidly scale.