Skip to content

Container Orchestration in PaaS Context

How PCF Relates to Kubernetes

Kubernetes: Container Orchestrator

Kubernetes is the system that: - Schedules containers on nodes - Manages networking between containers - Handles storage volumes - Implements service discovery - Provides autoscaling

You describe desired state → Kubernetes makes it happen.

PCF: Application Platform

PCF runs on top of Kubernetes (in cf4k8s) or implements similar orchestration concepts: - Schedules applications on infrastructure - Manages networking through routes - Binds services to applications - Implements service discovery through service brokers - Provides autoscaling

You deploy applications → PCF (using Kubernetes) makes it happen.

Conceptual Mapping

Kubernetes Concept PCF Equivalent Purpose
Pod Container/Instance Smallest deployable unit
Deployment Application Manage replicas
Service Route Network access
Namespace Organization/Space Multi-tenancy
ConfigMap Environment variables App configuration
Secret Credentials/Service binding Sensitive data
Ingress Route External access
StatefulSet Service backing resource Stateful workloads

Key Difference: Abstraction Level

graph TD
    A["Your App"] 

    A -->|PCF Abstraction| B["cf push"]
    B --> C["Platform handles everything"]
    C --> D["Results: Running app"]

    A -->|Kubernetes| E["Write YAML"]
    E --> F["Deployment<br/>Service<br/>ConfigMap<br/>etc"]
    F --> G["Results: Running containers"]

PCF: Higher Abstraction

  • Less control, but less complexity
  • Conventions over configuration
  • Built-in opinions (good defaults)
  • Faster time to production

Kubernetes: Lower Abstraction

  • Full flexibility and control
  • Configuration-heavy
  • You build the opinions
  • Longer time to production

Why cf4k8s?

cf4k8s runs Cloud Foundry on Kubernetes, giving you:

  • Best of both worlds: PCF developer experience + Kubernetes infrastructure
  • Multi-cloud: Run on any Kubernetes cluster
  • Local development: Run on Kind or Docker Desktop
  • Enterprise ready: Use existing Kubernetes infrastructure

cf4k8s = Cloud Foundry abstractions + Kubernetes orchestration

Container Concepts Remain the Same

Regardless of IaaS/Kubernetes/PCF:

  • Applications run in containers
  • Containers are stateless (preferably)
  • Data lives in services (databases, caches)
  • Horizontal scaling by running copies
  • Health checks determine restart policy
  • Networking is critical

For Your Learning

Since you know Kubernetes or container concepts:

  1. ✅ Skip container basics
  2. ✅ Focus on PCF-specific abstractions
  3. ✅ Understand why PCF makes different choices
  4. ✅ Learn when to use PCF vs bare Kubernetes

Next: PaaS Model