Development for Clouds

Approaches and Tools

Conventional approach with separate dev and ops:

• Long feedback cycles
• Source-target incompatibilities
• Slow upgrade of dev envs
• Slow adaption of new cloud-native frameworks
• Inconsistencies
• Poor transaction support or complex deployments

Development in and on the Cloud - shorten time to first dev cycle:

• Always-online requirement alleviated with distributed sync (git)
• Higher risk of vendor lock-in thus results may not be portable
• Dev/ops separation of concerns lifted

New trends: Immutable infra, NoOps, DevSecOps, Web IDEs

Models

Service Description Languages

• WSDL
• WADL
• USDL
• RAML
• CPL: Cloud Platform Language
• ABS: Abstract Behavioural Specification
• OpenAPI/Swagger
• Jolie: Language for microservices that allows you to centrally manage interfaces
• Ballerina: Another language for cloud stuff with a focus on http services

Top-down: generate code from description Bottom-up: generate description from code

Publishing

Bundle for deployment

• Service implementation
• Service description
• Service configuration
• Service contract template
• Composition description
• Client package

Deployment pipeline: Local -> Dev -> Test -> QA -> Prod

Service registration: Enabler for reuse (AWS API Gateway, Azure API Management)