Why should you care?

Banks and inventory systems depend on these guarantees when many users write at once.

See it live — copy this example

Run in SQL Server Management Studio (SSMS) or Azure Data Studio.

USE DataVerse;
-- Atomicity demo: both succeed or neither
BEGIN TRAN;
UPDATE dbo.Accounts SET Balance = Balance - 50 WHERE AccountId = 1;
UPDATE dbo.Accounts SET Balance = Balance + 50 WHERE AccountId = 2;
-- ROLLBACK;  -- uncomment to undo both
COMMIT;
SELECT AccountId, Balance FROM dbo.Accounts WHERE AccountId IN (1,2);

What happened?

• Atomicity is visible when ROLLBACK undoes both updates.
• Consistency is your CHECKs/FKs still holding.
• Isolation and durability show up under concurrency and restart.

Practice next

1. Commit a transfer and verify both balances.
2. Repeat with ROLLBACK and verify neither change remains.
3. Restart SQL Server service in lab and confirm committed data survived (durability).
4. Violate CHECK on purpose inside a tran and watch rollback.
5. Write a one-line definition of each ACID letter.

Remember

A = all or nothing. C = rules remain true. I = concurrency behavior; D = committed data survives.

Ledger trust

DataVerse banking markets ACID transfers to auditors.

Outcome: Committed postings survive crashes; partial postings do not.