Clusterless - Utilizing a Cluster of Clusters Seamlessly
Talkat KubeCon + CloudNativeCon North America 2021 · on Thursday, 2021-10-14· at 16:30· Concourse Hall 152 + Online·
summary
Design treats multiple clusters as Pod resource providers and preserves no new APIs by scheduling at pod/namespace granularity using virtual cluster primitives. Implementation extends virtual cluster with selective virtual cluster sinkers and a first‑level scheduler that annotates namespaces with placements, then dispatches pod slices to worker clusters; prototype runs a tenant control plane plus two worker clusters and achieves claimed O(n) scheduling throughput. Limitations: manual rescheduling, weak heterogeneity/HA support.
AI-written summary · about
chapters (9)
- 0:00Intro
- 0:25Why multicluster?
- 2:54Challenges
- 5:19Clusterless
- 6:56Cluster abstraction
- 8:07Pod vs. Workload
- 11:00Scheduling
- 14:00Multitenancy
- 15:50Interface
related talks
37mTenant Autonomy & Isolation In Multi-Tenant Kubernetes Clusters - A Comprehensive Guide
51mMulti-tenant Kubernetes
32mFlexible Kubernetes Multitenancy with vCluster
23mVirtual Kubernetes Clusters: A New Approach to Multi-tenancy
32mClusters wla Namespaces? Kifach blan Multitenancy f Kubernetes
38mDynamic Pod Resource Boundary Adjustment in Web Scale Clusters
05mPromoting Common Standards Without Sacrificing Tenant Autonomy
37mVirtual Clusters in the Real World: Lessons from k3k’s Journey to Production