Kubernetes Engine in comparison to Cloud Foundry

This guide compares two runtime services from STACKIT: STACKIT Cloud Foundry (SCF) and STACKIT Kubernetes Engine (SKE). We want to make sure you have reliable runtimes as part of the STACKIT cloud, so you can build and run your applications effectively.

We’ll look at different features to help you decide which runtime is best for your needs. These features include:

How the runtime is built (architecture).
How you set it up (deployment model).
How easily it can handle increased demand (scaling).
How traffic is directed to your applications (routing).
How well it works with other tools (ecosystem integration).
How much control you have (flexibility).
How you track performance and errors (logging and monitoring).
How resources are used (resource management).
How you manage your applications over time (application lifecycle management).
What skills your developers need (developer profile requirements).
How secure it is (security features).

By comparing these features, we’ll show you the strengths of each runtime and what types of applications they work best for. Do you prefer a platform that handles the underlying infrastructure for you, or do you need more control over your containers? This guide will help you choose the best STACKIT runtime service for your projects.

To dive deeper into each runtime service, visit the following links:

STACKIT Kubernetes Engine

Overview

STACKIT Kubernetes Engine is a managed Kubernetes service that provides fine-grained control over infrastructure and flexibility in managing containerized applications. It is suitable for teams comfortable with these technologies:

Kubernetes operates with a central, highly available, and redundant control plane to coordinate worker nodes within your cluster.
It runs containerized applications on worker nodes, providing the infrastructure needed for application execution.
Kubernetes enables declarative application operations, allowing you to define desired states and manage applications resiliently without human intervention.
Kubernetes facilitates and controls external access to services running within your cluster.
It automatically scales your application deployments based on demand, managing both the underlying infrastructure (nodes) and the application instances (pods).
The service is supported by a large, active community and an ecosystem of tools and services, ensuring continuous innovation and support.

By leveraging STACKIT Kubernetes Engine, you gain fine-grained control over infrastructure and benefit from advanced workload orchestration capabilities, allowing you to deploy, run, and maintain your applications with high reliability, optimal performance, and extensive customization options.

Target teams

Which individuals or teams benefit most from STACKIT Kubernetes Engine?

Site reliability engineers (SREs) requiring fine-grained control over reliability and scaling.
Platform engineers managing and customizing complex containerized environments.
Cloud architects designing multi-cloud infrastructures, ensuring portability across different cloud environments to avoid vendor lock-in.
DevOps teams seeking advanced automation and CI/CD integration with highly customizable workflows.
IT managers requiring flexible runtime services with extensive customization options.

Organization types

Which organizations benefit most from STACKIT Kubernetes Engine?

Enterprises with IT staff experienced in managing complex IT infrastructures.
Regulated entities requiring strict compliance and security controls with extensive customization capabilities.
Companies running distributed applications with multiple interconnected services and sophisticated orchestration needs.
Companies providing managed cloud services, such as data services or messaging services.

Typical use cases

Which use cases and scenarios are best supported by STACKIT Kubernetes Engine?

Containerized, highly scalable web applications with a microservice architecture.
Stateful services, such as database services, utilizing advanced Kubernetes operation concepts like Operators.
IoT applications using protocols other than HTTPS, such as AMQP and MQTT.
Containerized commercial off-the-shelf (COTS) applications running in multi-cloud environments.
AI and ML use cases requiring GPU usage, such as model training.
Solutions utilizing open-source components from the extensive CNCF ecosystem.

Cost considerations

This section explains the cost model, including a sample calculation for a 2-tier application deployed redundantly on STACKIT Kubernetes Engine.

Billing Model: Pay per use.
Billing Cost: €74.63 per cluster control plane per month; additional Compute Engine costs based on configured worker nodes per cluster.
Price list.
Cost calculator.

The application for the sample use case consists of these components:

Frontend
Backend
Application specific datastore
GitOps Tooling

Component	Node Type	Number of Worker Nodes	Monthly Avg. Hours	Total Cost
Cluster	g2i.4 CPU cores: 4 RAM: 16 GB Storage: 50 GB	2	720	€315.78
Control Plane	---	---	720	€74.63

STACKIT Cloud Foundry

Overview

STACKIT Cloud Foundry is a platform-as-a-service (PaaS) that abstracts infrastructure management, making it ideal for developers skilled in 12-Factor apps who prefer focusing on code:

Simplify deployment by streamlining the build process. Automatically package source code into containers and push them directly to the platform, enabling the signature cf push experience.
Use the roles and rights management system and multi-tenancy environment to support shared use by many customers.
Handle domain assignment, load balancing, and manage network configurations automatically.
Utilize the built-in service marketplace to create and bind data services to applications easily.
Use the included app autoscaler to adjust resources based on load, ensuring seamless application scaling.

STACKIT Cloud Foundry manages most operational responsibilities, enabling you to focus on delivering value through your applications.

Target teams

Which individuals or teams benefit most from STACKIT Cloud Foundry?

Application and full-stack developers who prioritize rapid development and focus on code rather than infrastructure.
DevOps engineers aiming to minimize infrastructure overhead in automated testing and deployment pipelines.
Teams with limited infrastructure expertise that need a managed platform experience.
IT managers looking for cost-effective cloud solutions with predictable spending.

Organization types

Which organizations benefit most from STACKIT Cloud Foundry?

Companies focusing on quick market entry, rapid prototyping, and application development.
Small and medium-sized enterprises (SMEs) seeking cost-effective scalability.
Organizations prioritizing developer productivity over infrastructure customization.

Typical use cases

Which use cases and scenarios are best supported by STACKIT Cloud Foundry?

Stateless web applications with a microservice architecture that require no more than 32 GB of memory per instance and automatically scale based on usage.
Containerized applications that serve requests via HTTPS or WebSockets.
Development of business applications focused on implementing pure business logic.
API services for mobile applications, other services, or third-party integrations.
Tasks that run scheduled or event-driven and consume no more than 32 GB of memory.

Cost considerations

This section explains the cost model, including a sample calculation for a 2-tier application deployed redundantly on STACKIT Cloud Foundry.

Billing Model: Pay per use.
Billing Cost: €0.03 per MB RAM per month.
Price list.
Cost calculator.

Sample use case:

Component	Used RAM	Number of Instances	Cost per MB RAM Used	Monthly Avg. Hours	Total Cost
Frontend	200 MB	2	€0.03	720	€12.00
Backend	600 MB	2	€0.03	720	€36.00

Choosing the right runtime service for your use case

Organizations should consider their team composition, technical requirements, operational model, and management preferences to select the most suitable runtime service.

Product details

This section provides an overview of the two STACKIT runtime services in comparison, serving as a quick reference.

Category	Criteria	STACKIT Cloud Foundry	STACKIT Kubernetes Engine
General	Architecture	Application-centric Focus on 12-factor apps and business application development Abstracts infrastructure HTTPS/web apps	Container-centric Platform to build platforms Provides infrastructure control and usage of GPUs Supports all kinds of protocols (e.g. IoT protocols)
	Documentation	SCF documentation	SKE documentation
	Pricing	Pay-per-use, MB RAM/hour (price list)	Pay-per-use, control plane + worker nodes depending on node types (price list), calculated per hour
Skills	Developer profile	Developers skilled in 12-factor applications Business application developers	DevOps engineers Developers familiar with the CNCF ecosystem
	Learning curve	Easy for developers, `cf push` experience	Requires infrastructure, Kubernetes and CI/CD knowledge
	Ecosystem	Cloud Foundry foundation	CNCF ecosystem
Availability	SLA	SCF Service Level Agreement	SKE Service Level Agreement
	Region availability	Available in all STACKIT regions	Available in all STACKIT regions
	High availability (multi AZ)	Supported in 3 availability zones	Supported in 3 availability zones
Operations & support	Release frequency	Bi-weekly, with ad-hoc patches for hotfixes	Bi-weekly, with ad-hoc patches for hotfixes
Operations & support	Planned downtimes / mandatory updates	No planned downtimes for multi-instance apps	Mandatory cluster updates when expired Kubernetes or OS versions are used, see documentation Updates executed as rolling updates
Scalability	Scaling	Horizontal autoscaling of application instances, minimum 1 instance running	Horizontal node autoscaling, minimum 1 node running Horizontal pod autoscaling
	Time for scaling	Instance scaling in range of seconds	Node autoscaling in range of minutes Pod autoscaling in range of seconds
	Maximum scaling limits	Unlimited number of app instances	1000 nodes per cluster
	Max. memory for workload	Max. 32 GB per app instance Max. 1.95 TB RAM per CF organization	Depends on selected node flavors
	Resource management	Automatic	Manual
Technical capabilities	Open-source based	Yes (GitHub)	Yes (GitHub)
	Flexibility	Opinionated approach	Customizable workflows, flexible orchestration
	Routing	Integrated HTTP routing	Allows flexible usage of ingress controllers for different protocols, such as HTTP, TCP, UDP, WebSockets, gRPC
	Logging and monitoring	Built-in tools	External tools
	Application lifecycle	Automated management	Manual management
	CI/CD	1-click deployment via cf push and buildpacks	Custom CI/CD pipeline for image build and deployment
	Deployable artifacts	OCI images	OCI images, Helm charts
	Supported languages	System buildpacks support Staticfile, Java, Ruby, .NET, Node.js, Go, Python, PHP, Binary, NGINX	Not limited to any specific programming languages
	Job/task support	Yes	Yes
	Available in STACKIT SNA	No	Yes
	Supports custom domains	Yes	Yes