Data and State Patterns

Serverless systems are stateless at the compute layer, not at the business layer. The real architecture lives in how uploads are stored, how databases are queried, how workflow progress is externalized, and how read performance is improved without turning a simple system into a fragile one.

Read this chapter in order. It starts with object storage as the default large-payload and artifact pattern, then moves into database access choices, durable state externalization, and finally cache or read-model design. Across all four lessons, the recurring question is the same: what state should live in a durable system, and what should never be trusted to function memory?

In this section

• Object Storage Patterns
  Describe the common role of object storage for uploads, static content, artifacts, snapshots, archival data, and event triggers. Explain where this pattern is especially powerful.
• Database Access Patterns
  Cover function interaction with relational databases, key-value stores, document stores, and analytical stores. Explain connection management, pooling concerns, and throughput patterns.
• State Externalization and State Stores
  Explain how temporary process state, checkpoints, locks, workflow progress, and derived state are moved into durable systems rather than kept in function memory.
• Caching and Materialized Views
  Describe caching patterns, precomputed views, TTL-based lookups, and read-optimization strategies that improve performance without breaking serverless simplicity.