Containers in the Microsoft ecosystem
Download as PPTX, PDF2 likes195 views
This document discusses containers and how they can be used to run applications in Microsoft Azure. It explains that containers package an application and its dependencies to run consistently across development, test, and production environments. Containers allow developers to package applications once and run them anywhere using the same tooling for both Linux and Windows. The document also outlines how containers can be orchestrated in Azure using services like Container Service, Service Fabric, Kubernetes, and Swarm.
Containers in the Microsoft ecosystem
- 2. Containers
- 4. Server Host OS Hypervisor Server Host OS Docker Engine Guest OS Guest OS Guest OS Bins/Libs Bins/Libs Bins/Libs App A App A’ App B Bins/Libs Bins/Libs AppA AppA’ AppB AppB’ AppB AppB’ AppB AppB’ Guest OS Bins/Libs App B’
- 6. Running containers in Microsoft Azure
- 12. Azure Container Service SwarmDC/OS ACS: The best place to run your containers Kubernetes
- 16. Real DevOps with Containers
- 17. Package an application and its dependencies Everything-as-code Consistency across development, test, & production Package once, run everywhere Same tooling for both Linux & Windows
- 18. Cloud Microsoft Azure On premises Third party Frameworks & Langages .Net PHP C++ JavaRuby Win32 Go Perl Python JavaScript Node Virtual Machines Containers Orchestrator Service fabric, Swarm, Kubernetes, DC/OS, Red Hat Open Shift…
- 19. Devs Develop the application Build a Docker image Push the image into a registry Ops Pull the image from a registry Composition, Configuration as Code, Execution Monitoring
- 20. … … … … … … Container Service Service Fabric Batch App Services Coming soon Build/CI, Integrate, Test 3. 1. Outer-Loop Push Code Production environments Run, Manage 5. 6. Code Run Validate Debug Inner-Loop CD, Deploy 4. Visual Studio Team Services Source Code Control (SCC) 2. Container Registry Azure
- 25. Containers = Operating system virtualization ContainerContainerContainer Traditional virtual machines = hardware virtualization VM VM VM Windows Server containers No different from Linux containers Container Container Container Hyper-V containers Isolation plus performance Container Container Container
- 28. Docker Platform Specific Platform Independent Linux Control Groups cgroups Namespaces Pid, net, ipc, mnt, uts Layer Capabilities Union Filesystems: AUFS, btrfs, vfs, zfs*,DeviceMapper Other OS Functionality Containerd + runC Docker Engine REST Interface libcontainerd graphlibnetwork plugins Windows Control Groups Job objects Namespaces Object Namespace, Process Table, Networking Layer Capabilities Registry, Union like filesystem extensions Other OS Functionality Compute Services Docker Client Docker SwarmDocker Compose Docker Registry
Editor's Notes
- #12: By creating an Azure Container Service you are creating two things, a set of agents on which your application will run and a set of masters which will manage and orchestrate your application. You will have one or more pools of machines called an Azure Container Service agent cluster (only one pool supported at time of writing). This can start as small as 1 machine and grow to thousands of machines. It is fully elastic, and with no single point of failure. All the machines in this cluster are identical. They are merely compute and associated resources on which our container application will run. You will also have 1, 3 or 5 machines called and Azure Container Service master cluster. This group is not elastic. Generally you will choose 1, 3 or 5 machines depending on the level of redundancy you require. With one master you run the risk of losing the ability to schedule new containers should that master go down, although deployed containers will continue to run if this happens and most state, if not all state, will be recovered when the master is restarted. With 3 and 5 masters you add additional availability to your service.
- #26: VM’s include application, necessary binaries and libs and an entire guest OS – all of these add up to tens of GB’s in size Containers include the app and all its dependencies, but share kernel with other containers. Run as isolated process on the host OS. Not tied to any infra , docker containers run on any computer, on any infrastructure and on any cloud. Containers do not provide sufficient isolation for many workloads. Hyper-V containers (Windows only) provide the same level of isolation has Hyper-V VMs but with minimal additional overhead compared to containers. The container itself is the same binary object whether you run as a Hyper-V container or a normal container.
- #29: Linux containers have been around for many years and consist of cgroups, namespaces, layers and a few other bits Docker built an engine that provides an API on top of these kernel technologies. They also built a bunch of tools that make it easy to work with containers via this API. For Windows Server Containers the Windows team added a “Computer Services” layer that provides similar features in the Windows Kernel. They then worked with Docker to port the open source Docker Engine to Windows. The API remains unchanged and thus all the existing tooling works across both Windows and Linux containers.