Mastering Cloud

Computing Paradigms

Computing paradigms (___ Computing)

• Distributed

  • Client-Server
  • Federated
  • Decentralised
  • Disconnected
• Service
• Cloud
• Utility
• Serverless

Application-level paradigms

• Cloud-native
• Serverless

Services

Technical characteristics

• Visibility: Well-defined, unified description (declarative, interactive)
• Interaction: Well-defined, unified interface/protocol (message exchange patterns, layer)
• Effect: Invocation semantics (idempotent, reversible on system, irreversible in real-world)
• Encapsulation: Encapsulate software and resources for re-use
• Interoperability: Loose coupling (late/dynamic binding)

Lifecycle characteristics

• Find (search, rank, select)
• Bind (signup, subscribe, pay, use)

Service
├── private
├── public
│   ├── non-profit
│   ├── commercial
│   │   ├── single-user
│   │   │   ├── standardised
│   │   │   ├── proprietary
│   │   ├── multi-user

Ecosystem

Service-oriented architecture

• Service orchestration or choreography
• Services, clients, brokers, marketplaces, messaging systems
• Service platforms, development tools

Web services (HTTP)

• RESTful, resource-centric (CRUD)
• SOAP, operation-centric (POST)

Content types and schemas

• Structured (JSON, YAML, XML)
• Plain text
• Binary documents
• Streams

Cloud services

• On-demand
• Pay-per-use
• Elasticity
• Robustness

Microservices

• Lightweight (e.g. single-function)
• Tangible & portable implementation

Scale

• Y-Axis: Scale by using microservices (functional decomposition)
• X-Axis: Scale by cloning services (horizontal duplication)
• Z-Axis: Scale by sharding databases (data partitioning)

Cloud-native & Server-less

Cloud-Native Computing: Computing paradigm optimised for modern distibuted systems

• ultra-scaling
• self-healing
• multi-tenant
• containerised
• dynamically-managed
• microservices-oriented
• resilient
• elastic
• virtualised
• loosely coupled
• abstracted
• adaptive

FaaS

• running functions in the cloud
• real pay-per-use