ABLENDER-finalpresentation
- 1. Deploying and Scaling Applications with Kubernetes Anna Blendermann 8/17/16
- 2. © 2015 LGS Innovations LLC 2 The Problem ● Containers have become incredibly prevalent in the software industry – How can we manage and scale this many Containers? ● Axios uses Containers to deploy and test applications – How can we manage, scale, and distribute our applications automatically and effectively?
- 3. © 2015 LGS Innovations LLC 3 Project Goals ● Become experts on Kubernetes concepts, structure, components, and network basics ● Deploy a test web application to the cluster and explore services and load-balancing features ● Breakdown a simple REDHAWK application into separate devices and deploy to the cluster ● Develop strategies and patterns for large scale REDHAWK cloud deployments
- 4. © 2015 LGS Innovations LLC 4 About Kubernetes ● Kubernetes, or K8s, is a open source system for managing Docker containers across a host of machines – Agile, application development and deployment – Automatic fault recovery and rolling updates – Portability across cloud and OS platforms
- 5. © 2015 LGS Innovations LLC 5 Docker Containers ● Docker – open source project that can deploy applications inside containers ● Container – virtual package that contains a piece of software and its dependencies ● Kubernetes orchestrates Docker containers
- 6. © 2015 LGS Innovations LLC 6 Kubernetes Terms ● Node – physical or virtual machine that manages pods on the cluster ● Pod – group of containers that share storage volumes, network data, and IP addresses ● Replica Controller – loop that maintains the number of running pods ● Deployment – object that creates and manages sets of replicated containers ● Service – abstraction that targets groups of pods through labels, ex: app=nginx
- 7. © 2015 LGS Innovations LLC 7 Nodes, Pods, & Containers Node Kubelet Kube-proxy Docker Pod Container
- 8. © 2015 LGS Innovations LLC 8 Kubernetes Structure Kubectl Kubelet Kube-proxy Docker API-Server Replica controller Kube-scheduler Service discovery Etcd
- 9. © 2015 LGS Innovations LLC 9 Work Accomplished ● Researched and recorded notes on Kubernetes concepts ● Created YAML files for pods, services, deployments, and replication controllers ● Deployed a simple REDHAWK application to the cluster ● Deployed multiple REDHAWK pods containing separate domains and devices ● Connected remote REDHAWK domains across the cluster using connections between pods ● Wrote bash scripts to automate cluster configuration and pod deployment
- 10. © 2015 LGS Innovations LLC 10 Technologies Used ● Docker ● Kubernetes ● Git ● Vagrant ● Python ● CentOS 6/7 ● Bash ● YAML
- 12. © 2015 LGS Innovations LLC 12 Challenges Faced ● Learning new software and languages from scratch ● Kubernetes GCE and Minikube clusters were difficult to replicate ● Understanding load balancing and service discovery ● Maintaining file consistency for storage volumes ● Keeping REDHAWK pod manifests updated
- 13. © 2015 LGS Innovations LLC 13 Concepts Learned ● Docker, Vagrant, Python, and YAML code ● Kubernetes architecture and history with Borg/Omega ● Deployment of applications using Kubernetes ● Cluster IP networking, load balancing, and services ● REDHAWK basics and command line execution ● More experience with Git and Bash tools
- 14. © 2015 LGS Innovations LLC 14 Future Work ● Launching REDHAWK on the Axios Engineering Kubernetes cluster with multiple remote domains ● Transfer data across the cluster using connections between devices and components ● Possible testing with other Axios container applications since pods are easy to create and deploy
- 15. Thank You