Transactions, ACID, BASE, and Consistency Tradeoffs
Understand how NoSQL systems handle correctness and why different databases offer different transactional guarantees.
Inside this chapter
- ACID and BASE in Practical Terms
- Not All NoSQL Systems Avoid Transactions
- Correctness Depends on the Workload
- Questions Advanced Engineers Ask
Series navigation
Study the chapters in order for the clearest path from NoSQL basics to advanced distributed design and production decision-making. Use the navigation at the bottom of each page to move through the full series.
ACID and BASE in Practical Terms
Relational databases are often associated with strong ACID guarantees. NoSQL systems are sometimes described with the term BASE, meaning more flexible consistency and availability tradeoffs in distributed environments. These are not slogans to memorize blindly. They describe real differences in behavior and design assumptions.
Not All NoSQL Systems Avoid Transactions
Some NoSQL systems now support multi-document or conditional transactions to varying degrees. But transaction support often comes with cost, limits, or different behavior than mature relational engines. Engineers should understand exactly what their system guarantees.
Correctness Depends on the Workload
A shopping cart can often tolerate looser consistency than a banking ledger. A recommendation feed can often accept eventual convergence where payroll cannot. Strong system design matches database guarantees to business risk.
Questions Advanced Engineers Ask
- What exact operation must be atomic?
- What stale data risk is acceptable?
- What happens during node or network failure?
- What guarantees are local, and which are distributed?