Hofls / cloud

Great alternative to on-premise computing

Quick overview

• Generic cloud computing platforms
  • AWS, Azure, GCP, DigitalOcean, Alibaba, Yandex
• Provides environment to run apps
  • Heroku
• Backend as a service
  • Firebase
• Hosting for static/serverless web apps
  • Netlify, Vercel
• Run your own cloud on your hardware
  • VMware vSphere

Cloud

• Cloud computing - on-demand delivery of IT resources over the Internet (e.g. computing power, data storage)
  • The main enabling technology for cloud computing is virtualization (separates a physical device from virtual devices)
• Types of computing:
  • On premises - you manage everything
  • IaaS (Infrastructure as a Service) - service manages networking, storage, servers, virtualization
    • AWS, DigitalOcean, Azure, GCE, Alibaba Cloud
  • PaaS (Platform as a Service) - IaaS + service manages OS, middleware, runtime
    • AWS Elastic Beanstalk, Heroku
  • SaaS (Software as a Service) - PaaS + service manages data and application (service manages everything)
    • Dropbox, Salesforce
  • FaaS (Function as a Service), aka Serverless - servers abstracted away.
    • Not suitable for complex apps. With lots of functions interacting with one another - complexity grows very quickly.
    • Azure functions,
  • BaaS (Backend as a Service) - provides auth, db, hosting, logs, analytics. You can focus on the frontend.
• Goals for the state of infrastructure:
  • Visibility - what services you are using, how you use them. Notifications when somebody makes changes. Detection of misconfigurations and incidents
  • Automation - automatically scale solutions, resolve incidents, rollback to previous configurations
  • Flexibility - how easy it is to make changes/improve your configurations
• Interfaces:
  • Web interface (e.g. AWS console) use cases:
    • Read-only usage (to understand state of the system)
    • When learning new service
    • Unique tasks (e.g. run EC2 isntance, check something, throw it away)
  • CLI (Command Line Interface):
    • Basic way to save and automate operations
  • API (Application Programming Interface):
    • Do not call API directly, better use SDK
  • SDK (Software Development Kit):
    • e.g. add SDK as dependency to build.gradle, write code that launches EC2 instances
• Levels
  • Architecture style -> Technology choices -> Application architecture -> Well-Architected framework
• Best practices:
  • Package your application and all dependencies in a Docker image
  • Introduce random instance termination during business hours (e.g. Chaos Monkey)

Cloud Comparison

• Service matrix
  • AWS vs Google Cloud vs Azure vs Openstack vs BYO (Open source)

Traditional (on premises) vs Modern (cloud)

On-premises	Cloud
Monolithic	Decomposed
Designed for predictable scalability	Designed for elastic scale
Relational database	Polyglot persistence (mix of storage technologies)
Synchronized processing	Asynchronous processing
Design to avoid failures (MTBF)	Design for failure (MTTR)
Occasional large updates	Frequent small updates
Manual management	Automated self-management
Snowflake servers	Immutable infrastructure