Online Transaction Processing

What is OLTP?

OLTP (Online Transaction Processing) is a database system designed to quickly record and process real-time transactions that occur in daily business operations, such as bank ATM withdrawals, e-commerce payments, and POS register operations. OLTP’s top priority is recording each transaction that occurs at that moment reliably, accurately, and quickly.

In a nutshell: “Right now, you’re buying something. That transaction is recorded reliably”—the heartbeat of business.

Key points:

• What it does: A system that records and processes real-time transactions reliably
• Why it’s needed: Business operations require transaction records to be both accurate and fast
• Who uses it: All business systems in every enterprise; indirectly, all customers

Why it matters

Modern society cannot function without OLTP. Bank transfers, shopping, salary payments, flight bookings—these activities can’t happen without reliable transaction recording.

OLTP’s importance lies in simultaneously achieving speed and reliability. Withdrawing 10,000 yen from an ATM must complete in seconds. But simultaneously, that withdrawal must be recorded accurately. If 10,000 yen is withdrawn without being recorded, the bank’s cash management fails. Conversely, if it’s recorded without being withdrawn, the customer loses money that doesn’t exist.

OLTP’s role is maintaining this balance while handling millions of simultaneous accesses. This is technically demanding; poor OLTP design destroys company credibility instantly.

How it works

Understanding OLTP is easier if you imagine old-fashioned ledger books. Banks kept notebooks for each account, handwriting deposits and withdrawals. Mistakes were unacceptable, so detailed rules governed each entry (debits in red, credits in black).

OLTP works the same way. When the database receives a command like “withdraw this amount from this customer’s account,” it follows this process:

First, check whether the withdrawal is possible. Is the balance sufficient? Is identity confirmed? Are there system problems? Next, execute the withdrawal. Reduce account balance and record the debit elsewhere. Finally, confirm the record. These three steps comprise a “transaction” and execute as “all-or-nothing”—either everything succeeds or everything fails.

This is crucial. If the system crashed mid-withdrawal, a normal system might end partially processed and corrupted. OLTP prevents this: either all steps in the transaction complete or none do—nothing falls in between. This is called “ACID properties” (Atomicity, Consistency, Isolation, Durability).

OLTP database design also balances speed and accuracy. Data is typically “normalized” (duplicates removed), minimizing update inconsistencies. Locks control simultaneous transactions.

Real-world use cases

Banking core systems

Bank core systems are OLTP’s classic example. Thousands of branches and millions of customer accounts access simultaneously. Each transaction (transfers, withdrawals, deposits) must complete in seconds. If processing took minutes, ATM customers couldn’t wait. OLTP systems handle this load through distributed processing across multiple servers, managing tens of thousands of simultaneous accesses.

Large e-commerce sales events

When an online store holds a “mega sale,” tens of thousands of orders flood in within minutes. Each order instantly completes: (1) inventory check, (2) charge processing, (3) shipping record creation. All must process accurately and simultaneously. Without OLTP, handling such demand spikes is impossible.

Airline reservation systems

Airline reservation systems experience simultaneous global access. “Tokyo-Osaka flight, tomorrow 10 AM, one seat available” is visible to hundreds simultaneously. Only one person actually purchases. The moment that one person completes purchase, the other 299 must see “sold out.” OLTP handles this competitive situation precisely.

Benefits and considerations

OLTP’s greatest benefit is providing reliable transaction support for core business operations. Real-time transaction recording also enables immediate customer service. Questions like “Where’s my delivery?” can be answered in real-time thanks to OLTP.

However, challenges exist. OLTP is optimized for “individual transactions,” so complex queries like “What were this month’s nationwide sales statistics?” don’t suit it. Running complex analytical queries on OLTP slows the system, degrading service for other users. For analytics, a separate system (OLAP) becomes necessary.

Additionally, OLTP demands 24/7/365 operation, creating large management costs. Server redundancy, regular backups, security measures—all require significant investment and personnel.

Related terms

• OLAP — Contrasts with OLTP; optimized for complex analysis
• Database Management System — OLTP is a primary DBMS use case
• Transaction Management — The core of OLTP that executes multiple operations as one
• Data Backup — Essential for maintaining OLTP reliability
• Business Systems — OLTP is the foundation for all enterprise business systems

Frequently asked questions

Q: Why separate OLTP and OLAP?

A: Design philosophies are completely different. OLTP specializes in “reliably processing one transaction now,” with data “normalized” (duplicates removed). OLAP specializes in “analyzing all historical data complexly,” with data “denormalized” (deliberately including duplicates). Running both in one system makes both slow, harming business.

Q: Is cloud database (like RDS) OLTP or OLAP?

A: It can be either, but it’s typically used for OLTP. Choose based on purpose—Amazon, for example, uses RDS (transactions) and Redshift (analytics) separately.