Databases 9 min read

An Introduction to the TPC-C Benchmark and Its Historical Development

This article introduces the TPC-C benchmark, its origin within the Transaction Processing Performance Council, the evolution from early standards like Debit Credit and TP1 to TPC-A, the structure of the TPC-C model with its nine tables and operations, and the subsequent development of newer standards such as TPC-E.

Tencent Database Technology
Tencent Database Technology
Tencent Database Technology
An Introduction to the TPC-C Benchmark and Its Historical Development

1. Citation

Recently, Chinese database OceanBase set new TPC-C world records for two consecutive years, drawing industry attention. Many readers may be unfamiliar with TPC-C, so this article explains the benchmark and its history.

2. Origin

The correct name is TPC-C, not TPCC. TPC stands for Transaction Processing Performance Council, a non‑profit organization that defines and oversees transaction‑processing performance standards.

In the 1980s, database vendors claimed superior online transaction processing (OLTP) capabilities, but no unified metric existed. In 1985, Jim Gray and colleagues published “A Measure of Transaction Processing Power,” introducing the Debit‑Credit (also called Debit Credit) benchmark, a system‑level standard that includes network and user‑interaction overhead.

IBM simultaneously proposed the TP1 (Transaction Process) benchmark for batch‑mode transaction performance. Without a governing body, vendors arbitrarily used Debit‑Credit or TP1, sometimes omitting requirements to improve results.

On August 10, 1988, eight companies formed the Transaction Processing Performance Council (TPC) to create and supervise standards, ending the earlier chaos.

3. TPC‑C Standard

TPC‑C is the “C” standard issued by the TPC organization; the hyphen is essential. It follows earlier standards TPC‑A (released in 1989) and TPC‑B. TPC‑A saw 33 companies adopt it, producing 115 system results, marking the first widely accepted benchmark.

As technology advanced, TPC‑A’s limitations became apparent, leading to the development of TPC‑C, published in 1992. TPC‑C models a wholesale‑goods transaction scenario and has been the dominant benchmark for over a decade, with the latest version (5.11.0) released in 2010.

4. TPC‑C Model

The TPC‑C model consists of nine tables that simulate a wholesale system:

Warehouse – records storage locations.

District – each warehouse has ten districts (1‑to‑10 relationship).

Customer – 3,000 customers per district.

Order – records each customer’s purchase.

New‑Order – a subset of Order for pending orders.

The benchmark defines five operations (new‑order, payment, order‑status, delivery, stock‑level) that together emulate a warehouse‑centric wholesale workflow. Performance is measured by tpmC (transactions per minute), derived from the number of new orders created per second.

Because TPC‑C is based on a wholesale model, it does not fully reflect modern B2B/B2C workloads. Consequently, the TPC introduced TPC‑E in 2007 to model securities‑market trading, and other standards such as TPC‑H exist for different scenarios.

5. Summary

We now have a basic understanding of the TPC committee and the benchmarks it has produced, including TPC‑A, TPC‑C, TPC‑H, and TPC‑E, as well as the current and retired standards.

Future articles will delve deeper into each TPC benchmark.

Benchmarkdatabase performancetransaction processingTPCTPC-CTPC-E
Tencent Database Technology
Written by

Tencent Database Technology

Tencent's Database R&D team supports internal services such as WeChat Pay, WeChat Red Packets, Tencent Advertising, and Tencent Music, and provides external support on Tencent Cloud for TencentDB products like CynosDB, CDB, and TDSQL. This public account aims to promote and share professional database knowledge, growing together with database enthusiasts.

0 followers
Reader feedback

How this landed with the community

login Sign in to like

Rate this article

Was this worth your time?

Sign in to rate
Discussion

0 Comments

Thoughtful readers leave field notes, pushback, and hard-won operational detail here.