Module 08: Cloud Architecture

Goal

Design reliable, secure, scalable, observable, and cost-aware cloud architectures for AI systems.

Core Areas

• Compute: VMs, containers, serverless functions, managed container platforms.
• Storage: object storage, file storage, block storage.
• Databases: relational databases, NoSQL, cache, vector databases.
• Networking: VPC/VNet, subnets, firewalls, load balancers, private endpoints, DNS.
• Identity: IAM, RBAC, managed identities, service accounts, key management.
• AI services: model deployment, hosted inference, search, monitoring, governance.
• Operations: logs, metrics, traces, alerts, dashboards.

Architecture Qualities

• Scalability.
• Reliability.
• Security.
• Cost efficiency.
• Performance.
• Operability.
• Maintainability.

Study Tasks

• Draw a reference architecture for a RAG platform.
• Draw a reference architecture for an agent platform.
• Map cloud services across at least two providers.
• Estimate monthly cost for a small production deployment.
• Define backup and disaster recovery strategy.
• Define monitoring and alerting strategy.
• Create an architecture decision record for provider selection.

Project

Build projects/07-production-cloud-ai-platform/.

Tips

• Choose one primary provider for depth, but learn equivalent concepts elsewhere.
• Start with architecture qualities before choosing services.
• Every external dependency needs an outage plan.
• Cost estimation is part of architecture, not finance paperwork.

Completion Checklist

• Can design compute, storage, database, and network layers.
• Can explain IAM and private networking.
• Can design for high availability.
• Can estimate cloud and model inference cost.
• Can document trade-offs.