mck8s: A Container Orchestration Platform for Geo-distributed Multi-cluster Environments
summary
mck8s is a multi-cluster Kubernetes orchestration platform that implements four controllers — a multi-cluster scheduler, a horizontal pod autoscaler, a multi-cluster rescheduler and a cloud cluster provisioner — and integrates Kubernetes Federation, Cluster API, Prometheus, Cilium and surf. It runs across a management cluster plus five regional workloads and an OpenStack cloud; demos show offloading and bursting of replicas (5 → 10 → 15 triggers automated cloud1 provisioning) and placement via best-fit, worst-fit and traffic-aware policies, with automated node scaling and cross-cluster routing working end to end.
AI-written summary · about
chapters (23)
- 0:00Start
- 1:02Evolution of Cloud Deployments
- 1:56Major Types of Geo-Distributed Deployments
- 2:40Emerging Fork Computing Paradigm
- 4:32Resource Management and Application Deployment Challenges
- 6:14Kubernetes Architecture
- 7:41The Kubernetes Federation
- 10:18Architecture of Mckate
- 12:59Multi-Cluster Scheduler
- 13:48Automated Resource-Based and Network-Based Policies
- 14:46Multi-Cluster Schedule Scheduler
- 16:05Manifest File for a Sample Multi-Cluster Deployment
- 17:53Horizontal Port Autoscaler
- 20:15Demo
- 21:42Prerequisites
- 23:46The Presence of the Cluster Api Resources
- 24:12Presence of the Openstack Cluster
- 27:38Deploy Multi-Clusters Applications
- 28:57Create a Cloud Provisioner
- 32:59Bursting
- 37:48Best Placement Policy
- 38:43Westfield Policy
- 39:33To Deploy the Multi-Cluster Service Controllers
related talks
33mMcK8s: Container Orchestration in Kubernetes Multi-Clusters
19mKeynote: Kubernetes Federation
29mMulticluster Kubernetes Management Made Easy With Open Cluster Management
24mSelf-Service Multi-Cloud K8s Platform at a Regulated Financial Enterprise
33mKubernetes-Driven Multi-Cloud Deployments: Maximizing Scalability and Flexibility
18mA Cloud-Native Tool for the mass
38mMulticluster Kubernetes, today and tomorrow
28mOverview Of Challenges And Solutions For Orchestrating Applications To Multiple DC And Edge Clusters
