DP-900 · 1 Describe core data concepts
Transactional workloads
Exam objective: Describe features of transactional workloads
Transactional processing, or OLTP, records individual business events quickly and reliably. ACID properties guarantee that each transaction is trustworthy.
Online transactional processing, or OLTP, is what most business applications do most of the time: record discrete events as they happen. A transaction could be a bank transfer, a retail payment or any other small unit of work that the organization needs to track accurately, sometimes millions of times a day.
An OLTP database has to support fast create, retrieve, update and delete operations while keeping every transaction trustworthy. Microsoft Learn describes that trustworthiness with four ACID properties:
| Property | What it guarantees |
|---|---|
| Atomicity | the transaction fully succeeds or fully fails |
| Consistency | the data moves between valid states only |
| Isolation | concurrent transactions do not see each other's partial work |
| Durability | a committed transaction survives even if the system goes down |
Take a transfer of funds between two accounts: debiting one account and crediting the other has to happen as a single unit, or not at all.
On the exam, look for OLTP, CRUD, ACID or a scenario describing high volume, time sensitive business transactions.
Key points
- A transaction is a small, discrete unit of work, such as moving money between two accounts.
- OLTP systems support CRUD operations: create, retrieve, update and delete records, optimized for both reads and writes.
- Atomicity means a transaction succeeds completely or fails completely, with no partial result left behind.
- Consistency means a transaction only moves the data from one valid state to another valid state.
- Isolation keeps concurrent transactions from seeing each other's half finished work, and durability means a committed transaction survives even a power loss.
Exam trap
Durability and isolation are easy to swap. Durability is about surviving a crash after a commit. Isolation is about concurrent transactions not interfering with each other while they run. Neither one is about recovering from a failed transaction, that is atomicity.
Check yourself
A bank transfer debits one account and credits another. The system must apply both changes together, or neither change at all. Which ACID property does this scenario illustrate?
Go deeper on Microsoft Learn
Checked against Microsoft Learn on October 1, 2026.