Cloud Native

Architecture for Elasticity and Resilience

Cloud Native is not simply moving servers to the cloud (Lift and Shift). It is an architectural approach in which applications are designed specifically to leverage the advantages of distributed cloud environments: elasticity, resilience, and fast iteration cycles.

At its core are containerisation (Docker), orchestration (Kubernetes), and the decomposition of applications into small, independently scalable services.

Anti-Patterns: Cloud Migration Without Value

• Lift and Shift of legacy systems: Moving a rigid monolith 1:1 onto a virtual machine in the cloud. The result: higher costs without gaining any of the benefits of cloud flexibility.
• Manual server maintenance: Managing cloud servers the same way as physical machines in a server room (applying patches, manual configuration), which prevents scaling.
• Vendor lock-in through proprietary services: Using vendor-specific features that make switching providers or running multi-cloud setups impossible.

Cloud-Native Architecture

1. Containerisation: Packaging code and all its dependencies into standardised containers for high portability across environments.
2. Kubernetes (K8s) Orchestration: Automated deployment, scaling, and management of the container fleet.
3. Microservices and APIs: Decomposing the application into small units that communicate via network interfaces (see API-First).
4. Stateless Design: Applications store no data locally, enabling instances to start, stop, and move quickly.
5. Observability: End-to-end telemetry (logs, metrics, traces) to understand the state of the complex distributed system at any time.

The Focus: Business Resilience

Well-designed cloud-native systems can repair themselves (Self-healing) and automatically adjust their capacity to actual demand, which can reduce costs and increase availability. These properties result from health checks, controllers, redundancy, and scaling policies; they are not automatic.

FAQ

Why do we need Kubernetes? Isn't it too complex?

It is a common standard for container orchestration, but not mandatory for every cloud-native system (serverless, managed PaaS, and container services are alternatives). Where it fits, it takes the work out of manually monitoring and scaling servers, and forms the foundation for a modern, automated IT operation.

Is Cloud Native more expensive than traditional hosting?

Infrastructure costs may increase, but operational costs drop significantly through automation. You also gain enormous speed in shipping new features.

References

• Wiggins, Adam The Twelve-Factor App. Methodological foundation for modern SaaS applications. (2012). 12factor.net
• CNCF Cloud Native Definition v1.1. Definition and ecosystem overview. github.com/cncf/toc/blob/main/DEFINITION.md
• CNCF Kubernetes Documentation. Reference for container orchestration. kubernetes.io/docs/home/