Container orchestration overview
Download as PPTX, PDF1 like1,401 views
The document provides an overview of container management, highlighting the significance of containers in cloud architecture and software deployment, as well as the challenges of managing multiple containers. It discusses various tools and systems for container orchestration, such as Kubernetes, Swarm, and ECS, and outlines the features and benefits of these systems. Additionally, it mentions training opportunities and job openings at Excella related to DevOps and container management.
Container orchestration overview
- 1. excella.com | @excellacoWyn Van Devanter | @wynv Overview of Container Management
- 2. excella.com | @excellaco Excella Announcements NEW: Advanced Certified ScrumMaster (A-CSM) Training Upcoming Training Certified ScrumMaster (CSM) August 2 – 3 | Arlington, VA Certified Scrum Product Owner (CSPO) August 9 – 10 | Arlington, VA Certified Kanban System Design (KMP I) August 27 – 28 | Arlington, VA Register for these upcoming classes at www.excella.com/training We’re hiring DevOps Engineer Senior DevOps Engineer DevOps Manager Cloud Architect More… www.excella.com/careers
- 3. excella.com | @excellaco Agenda Why Container Management? Quick Container Overview Clusters, Cloud Architecture & Containers Container Orchestration Tool Overview Container Management Systems and Beyond
- 4. excella.com | @excellaco Why container management? Building using containers is becoming the standard. Integral part of cloud architecture. Creates new ways to package, deploy and manage software Centralizes software management Fosters standardized deployment platform
- 5. excella.com | @excellaco Container-based Infrastructure Container Management System Orchestrator Container Engine Containerized Applications Security controls, image security scanning, centralized management tools, app lifecycle management, enterprise management Scheduling, communication, service discovery, load balancing, self- healing, rolling updates, pipeline management, federation, etc Runs containers Application packaged in a standard way Docker EE, OpenShift, control plane Cluster(s) managed by Kubernetes, Swarm, Mesos, Fleet Docker on Azure/AWS/VM App with React UI container, ASP.NET Core API container
- 6. excella.com | @excellaco What are containers? Isolate an application, its dependencies and resource use into a standard unit of execution & deployment Your application and everything it needs to run, with isolation benefits without the OS overhead More portable, less resource use, more shippable…
- 7. excella.com | @excellaco What are containers? ContainersVirtual Machines
- 8. excella.com | @excellaco Deploying Containers Must manage many pieces (containers) Must be able to find each other Immutable infrastructure, upgrading Distributing workload Scheduling scaling Recovery More…
- 9. excella.com | @excellaco All these containers • Apps in Linux Containers, running against a Linux kernel, on a Linux server(s) • Apps in Windows Containers, running against a Windows kernel, on a Windows server(s) • Servers in AWS, servers in Azure, in-house All containers, so all the same in some respects. Manage them all the same way, centrally.
- 10. excella.com | @excellaco Clusters • Backbone of container infrastructure. Typically for large scale, now container-based software deployment too. • Makes managing a pool of servers & their resources as simple as managing a single system • You don’t want to care much, about where containers run • With multiple container-based apps, a cluster is a way to standardized deployment
- 11. excella.com | @excellaco Cloud architecture • Immutable, identical infrastructure • Distributed, decoupled • APIs provide buffer between implementer and consumer • Load balancers provide buffer between instances of each service • Self-healing • Scalable • Declarative state • Self-service infrastructure for developers to consume over cloud vendor-specific APIs
- 12. excella.com | @excellaco What is Container Orchestration? • Tooling that makes managing container-based applications across a cluster(s) easier • An abstraction that simplifies tasks of building, deploying, and maintaining scalable distributed systems in the cloud • Decouples development teams from machines they’re using • Single platform for application deployment across clouds • Operationalized efficiency across the organization • Various distributed container management functionality
- 13. excella.com | @excellaco Container Orchestration • Tools called orchestrators / schedulers manages containers across cluster(s) • Kubernetes, Swarm, Mesos/Marathon, Nomad Managed: ECS, EKS, Fargate, AKS, GKS • Automates the distribution of applications across a cluster of machines, ensuring higher levels of utilization
- 14. excella.com | @excellaco Container Management Systems
- 15. excella.com | @excellaco Container Management needs • We don't have a straightforward way to manage the scaling, deployment and communication of multiple containers across machines. • If the server reboots or Docker restarts, the container and app will shut down and not come back up. • We also have to consider how we will monitor the health of the service and its logs, among other services deployed.
- 16. excella.com | @excellaco Container Orchestration • Helps make the pieces in a distributed container system work together • Scheduling - placing a container on the right node • Cross-node communication • Service discovery • Scaling • Load balancing • Self-healing (automatic restarts) • Rolling updates • Service/App management (organizing containers that comprise a service, deployments), blue/green & canary deployments • Storage management
- 17. excella.com | @excellaco Orchestrators Container Management System Orchestrator Container Engine Containerized Applications Security controls, image security scanning, centralized management tools, app lifecycle management, enterprise management Scheduling, communication, service discovery, load balancing, self- healing, rolling updates, pipeline management, federation, etc Runs containers Application packaged in a standard way Docker EE, OpenShift, control plane Cluster(s) managed by Kubernetes, Swarm, Mesos, Fleet Docker on Azure/AWS/VM App with React UI container, ASP.NET Core API container
- 19. excella.com | @excellaco Container Orchestration
- 20. excella.com | @excellaco Cluster of Nodes • Nodes (hosts) pool resources to form cluster • In a typical container-based environment, you would typically define that specific containers run on specified nodes, but that can be partially or fully managed. • The containers that comprise an application can move from node to node based on various conditions. • Any node can direct traffic to where the container(s) is being hosted.
- 21. excella.com | @excellaco Setting up an orchestrator • Number of nodes in a cluster • Specification & settings of the containers comprise your application or batch job • Resource usage • Other settings • CLI tools, etc. to create custom clusters, to SaaS-based cluster management
- 22. excella.com | @excellaco Tool Overview • Kubernetes • Swarm - native Docker API, defaults easy to set up • ECS – managed, AWS orchestrator (ACS) • EKS – managed, Kubernetes orchestrator (AKS) • Fargate – more infrastructure management • Mesos with Marathon/etc • Nomad: light; cluster management and scheduling only provided; rest with add-ins like Consul, Vault; operationally simple
- 23. excella.com | @excellaco Choosing a Container Management System • What kinds components are going into containers? • How do the components of the application talk to one another? • How is availability handled? • How is access control handled? • What sort of scalability is needed • How many applications am I planning to support? • Do I need to manage my own cluster(s)?
- 24. excella.com | @excellaco How the tools work • Create configuration file that specifies containers and other settings that make up a service • Tool places containers on nodes in cluster • Makes sure they are healthy • Can manage deployments from the tool
- 25. excella.com | @excellaco Some Concepts (Kubernetes) • Master - normally doesn't run containers, but the key Kubernetes services; nodes • Namespaces - logical grouping of cluster for use with multiple users or projects • Labels - K/V pair for categorizing objects such as pods • Pods - wraps container(s) • Deployments - manages the desired state, i.e. specify number of pods • Services - allows external pod communication • Volumes - share data between containers, persistent storage; beefier than Docker volumes, tied to life of
- 26. excella.com | @excellaco Kubernetes concepts • Can group together container images developed by different teams into a single deployable unit • Kubernetes services provide load balancing, naming, and discovery to isolate one microservice from another • Ingress objects provide an easy-to-use frontend that can combine multiple microservices into a single externalized API surface area.
- 27. excella.com | @excellaco Kubernetes for developers • Application-oriented container APIs • Separate developers from specific machines • Increase portability since developers are consuming a higher-level API that is implemented in terms of the specific cloud infrastructure APIs • Can abstract you from a particular cloud. For example, Kubernetes services know how to create load balancers on all major public clouds as well as several different private and physical infrastructures. • Minikube
- 28. excella.com | @excellaco Kubernetes tidbits • Google-spawned Kubernetes has quickly risen to dominate the container orchestration space • Heavyweights behind it including Red Hat, IBM • 71% of major enterprises use Kubernetes • More focused on applications so plumbing is more set up for that • It can be harder to set up for other cases. More configurable but more complex. • Composes applications via pods – i.e. 1-5 containers making a service
- 29. excella.com | @excellaco Kubernetes Source: Red Hat
- 30. excella.com | @excellaco Kubernetes Capabilities: • Scheduling • Cross-node communication • Service discovery • Scaling, autoscaling • Load balancing • Self-healing (automatic restarts) • Rolling updates
- 31. excella.com | @excellaco Kubernetes • Advanced networking makes multi-host clusters much more easy to manage, at the expensive of slightly more complexity • Pods facilitates groups of containers that can communicate with each other as though they are running within the same system. • Allow applications running within the pod to interact in the same way they would as though they were not running in containers but simply running on the same physical host.
- 32. excella.com | @excellaco Swarm • Swarm is more general purpose so setup decisions are left more open. But easy to get running. • Uses standard Docker API, so things like docker run work as usual • Tools like Compose can use Swarm without any changes.
- 33. excella.com | @excellaco Swarm Capabilities: • Scheduling • Cross-node communication • Service discovery • Scaling, autoscaling • Load balancing • Self-healing (automatic restarts) • Rolling updates
- 34. excella.com | @excellaco ECS • Managed AWS service launched in 2014 • Composes applications using: • Task definitions • Tasks • Services • Works with other AWS services such as ELB, elastic block storage, CloudWatch, VPC, CloudTrail • No support for infrastructure outside of EC2
- 35. excella.com | @excellaco ECS Capabilities: • Scheduling • Load balancing • Service Discovery • Service management - task definitions, tasks, services • Logging, Auditing • Storage
- 38. excella.com | @excellaco Control Plane Engineering Sits on top of the server nodes to help better management, uptime and performance across multiple clouds Creates a deployment that’s easily scaled Run the same system on multiple, hybrid clouds
- 39. excella.com | @excellaco Control Plane in Kubernetes
- 40. excella.com | @excellaco Container Management Systems Container Management System Orchestrator Container Engine Containerized Applications Security controls, image security scanning, centralized management tools, app lifecycle management, enterprise management Scheduling, communication, service discovery, load balancing, self- healing, rolling updates, pipeline management, federation, etc Runs containers Application packaged in a standard way Docker EE, OpenShift Cluster(s) managed by Kubernetes, Swarm, Mesos, Fleet Docker on Azure/AWS/VM App with React UI container, ASP.NET Core API container
- 41. excella.com | @excellaco Container Management Systems • Enterprise-grade (Docker Enterprise, OpenShift) • Additional functionality, enterprise features, etc • More security scanning • secured images • centralized access control & management • networking • pipeline features (i.e. easy multiple environment setup), standardization • CaaS/PaaS functionality • cloud-based services • support
- 44. excella.com | @excellaco Thank you! @wynv | wyn.vandevanter@excella.com • Getting Started with Docker, Self-Guided Workshop, https://github.com/excellalabs/docker-workshop-1 • Getting Started Deploying with Docker via Kubernetes, Self-Guided Workshop, https://github.com/excellalabs/docker-workshop-2 • Docker Swarm workshop, https://github.com/jpetazzo/container.training • Deploy ASP.NET Core app to Kubernetes on Google Kubernetes Engine, https://codelabs.developers.google.com/codelabs/cloud-kubernetes-aspnetcore