Cost effectiveness

• traditionally, costs associates were upfront, considered capex

• moving to cloud, it is considered opex

• so can be cost effective just showing this.

• pay less for services

• related to [[202404061425 Azure Cost Management|cost]]

how?

• plan and estimate: design and then use cost-estimation to figure out how much you will pay
  • get discounts where possible by using reserved instances for example. also for licenses, etc.
• optimization: design with cost optimization at the center.
  • PaaS is better than IaaS. choose service levels, etc based on usage requirements.
  • optimize consumption based pricing: compare prices for dedicated vs consumption based pricing for apps.
  • optimize HA: active-active or active only is better than active-passive
  • clean unused apps/services
• monitoring: related to above. monitor and figure out. and then resize/optimize.
• efficiency: related to above. ensure everything is utilized effectively. reduce waste.
  • examples: VM utilization (should not be free). storage utilization (data not required to be read often, not kept in archive) etc.

Operational excellence

• full visibility into how your applications is working

how?

• devops practices to ensure agility. ci/cd containers. or legacy. does not matter.
• effective monitoring - identify cost wastes and troubleshooting, etc.
• automation - reduce human errors. easier management.
• testing - reduce issues with releases.

Performance Efficiency

• matching available resources with the demands being put on the app

how?

• scaling up and out
  • up/vertical - make something larger
  • out/horizontal - add similar resources
  • ideally, use autoscaling
• optimize network
  • add message layer, so that requests keep flowing
• optimize storage
  • partitioning to access/manage data differently
  • caching
    • placing static content closer to users, i.e. CDNs
• identify poorly performing code or bottlenecks
  • performance monitoring

Reliability

• how to handle failures
  • hardware failures
  • data loss
• including ha in our design

how?

• build ha systems dependent of the sla we are committing
• recoverability/ once the below are defined, we can design accordingly
  • related to [[202404071556 Disaster Recovery|Disaster Recovery]]
  • [[202404081931 Recovery Point Objective|RPO]]
  • [[202404081933 Recovery Time Objective|RTO]]

Security

• protecting data
• [[202404051739 Governance Overview|azure governance]] i.e. shared responsibility model depending on what you’re using

how?

• increase security in layers/ so that attacker has to do more work
  • data - exposing encryption key or using weak encryption can cause issues
  • apps - code injection and execution - eg sql injection and cross site scripting
  • vm/compute - malwares
  • networking - unnecessarily open ports are a problem
  • perimeter - ddos attacks
  • policies and access - exposure of creds. limit access. monitoring to see where logins are coming from.
  • physical - security badge stealing. door drafting, etc.

---

references:

MS Learn - Well architected Framework MS Learn Docs - WAF