An access pattern is a repeatable way an application reads from or writes to storage. A schema describes what data looks like; access patterns describe how it moves. Choosing a database with those patterns in mind is more useful than shopping by feature list alone. section 8.2.1
Map the traffic your application actually expects:
- What shapes do queries take, and how do they find data? section 8.2.1
- How many reads occur compared with writes? How large are the payloads? section 8.2.1
- How much latency can the application tolerate? section 8.2.1
Orders and payments
A request for active orders placed by customers in Asia over the past 30 days, alongside their payment history, crosses several domains. A relational database can answer this kind of complex request using SQL joins. section 8.2.1
Start with a relational database for typical business data, but let evidence about the actual workload guide the choice—not hunches. section 8.2.1




