Lesson 148

Transactions, ACID & NoSQL

ACID · Isolation · CAP · NoSQL

1:00

How databases stay correct under concurrency and crashes — the ACID contract, isolation levels, the CAP theorem, and when to choose NoSQL over relational.

By the end, you can

  • Explain what a database transaction is and why it is defined as all-or-nothing.
  • Define each of the four ACID properties and give a one-sentence description of each.
  • Identify dirty reads, non-repeatable reads, phantom reads, and lost updates and state which isolation level prevents each.
  • Order the four standard isolation levels from weakest to strongest and state what each prevents.
  • Contrast Two-Phase Locking and MVCC as concurrency-control strategies.
  • Explain the Write-Ahead Log pattern and how it delivers both Atomicity and Durability.
  • Name the four NoSQL families, give a canonical example database for each, and describe the workload each fits best.
  • State the CAP theorem and explain the CP vs AP trade-off during a network partition.
  • Distinguish ACID (strong, immediate consistency) from BASE (eventual consistency) and identify when each is appropriate.
Up next in Networking, Distributed Systems & Databases
Questions or feedback?