Skip to main content

Best Practices

General Principles​

  • Usability and Simplicity
    • Design the architecture to be easy to use, so that it can be built, maintained and adopted by a wide range of users.
  • Scalability
    • Ensure that the architecture is designed to scale horizontally and vertically to accommodate the expected growth in traffic and data.
  • Resilience
    • Implement mechanisms to ensure that the system can handle failures, such as load balancing and fault tolerance.
  • Security
    • Incorporate security measures to protect sensitive data and prevent unauthorized access.
  • Performance
    • Optimize the architecture for high performance by considering factors such as network latency, data access patterns, and caching strategies.
  • Modularity
    • Design the architecture to be modular, with loosely coupled components that can be easily replaced or upgraded.
  • Automation
    • Automate as much as possible, such as provisioning, deployment, and monitoring to reduce manual errors and increase efficiency.
  • Monitoring
    • Monitor the system to detect and diagnose issues, and to gather metrics for capacity planning and performance optimization.
  • Cost optimization
    • Consider the cost of the architecture, both in terms of initial investment and ongoing maintenance, and optimize for cost where possible.
  • Flexibility
    • Design the architecture to be flexible, so that it can easily adapt to changing requirements and technologies.
  • Continuous integration and delivery
    • Implement continuous integration and delivery (CI/CD) practices to streamline the development and deployment process.

Google Cloud Principles​

Google Cloud Architecture Framework

Core principles of system design​

  • Document everything
  • Simplify your design and use fully managed services
  • Decouple your architecture
  • Use a stateless architecture

Geography​

  • Deploy over multiple regions
  • Select regions based on geographic proximity
  • Select regions based on available services
  • Choose regions to support compliance
  • Compare pricing of major resources
  • Use Cloud Load Balancing to serve global users
  • Use the Cloud Region Picker to support sustainability

Manage cloud resources​

https://cloud.google.com/architecture/framework/system-design/resource-management

  • Use a simple folder structure for labels/tags
  • Use folders and projects to reflect data governance policies
  • Use tags and labels at the outset of your project
  • Assign labels to support cost and billing reporting
  • Avoid creating large numbers of labels
  • Avoid adding sensitive information to labels
  • Anonymize information in project names
  • Don't include attributes that can change in the future, for example, a team name or technology.
  • Apply tags to model business dimensions
  • Apply tags to model business dimensions

Organizational policies​

  • Establish project naming conventions
  • Automate project creation
  • Audit your systems regularly
  • Configure projects consistently
  • Decouple and isolate workloads or environments
  • Enforce billing isolation
  • Use the Organization Policy Service to control resources
  • Use the Organization Policy Service to comply with regulatory policies
  • Limit resource sharing based on domain
  • Disable service account and key creation
  • Restrict the physical location of new resources

Choose and manage compute​

  • Choose a compute platform
    • Serverless
    • Containers
    • Instances
  • Choose a compute migration approach
    • Lift and shift
    • Upgrade to containers

Designing workloads​

  • Evaluate serverless options for simple logic
  • Decouple your applications to be stateless
  • Use caching logic when you decouple architectures
  • Use live migrations to facilitate upgrades

Scaling workloads​

  • Use startup and shutdown scripts
  • Use MIGs to support VM management
  • Use pod autoscalers to scale your GKE workloads
  • Distribute application traffic
  • Automate compute creation and management

Managing operations​

  • Use Google-supplied public images
  • Use snapshots for instance backups
  • Use a machine image to enable VM instance creation
  • Evaluate built-in migration tools
  • Use virtual disk import for customized operating systems
  • evaluate whether your application is listed under a supported vendor.

Capacity, reservations, and isolation​

  • Use committed-use discounts to reduce costs
  • Choose machine types to support cost and performance
  • Use sole-tenant nodes to support compliance needs
  • Use reservations to ensure resource availability

VPC networking​

  • Consider VPC network design early
  • Start with a single VPC network
  • Keep VPC network topology simple
  • Use VPC networks in custom mode
  • Inter-VPC connectivity
  • Use simple naming conventions
  • Use connectivity tests to verify network security

Cloud Services Best Practices​

AWS services​

Microsoft Azure services​

Google Cloud services​

Architecture Topic References​

Resilieance​