Comparison with TPC Benchmarks

Laboratory Home
Test Results
This Site
Email List
Neal Nelson
& Assoc. Home

The following comments identify areas of similarity and areas of difference for the Neal Nelson Transaction Benchmark and the Transaction Processing Council tests.


Transaction Processing Council

Neal Nelson Transaction Benchmark

Operating System Software
     Permits the use of many different operating systems for a test. Requires the use of SUSE Linux Enterprise Server version 10.
System Tunables
     Allows the person running the test (normally a computer vendor) to set the tunables in whatever way they want. The settings are included in the test's full report. Sets the tunables the same for all tested systems.
Application Software
     Defines the data tables, transaction types, transaction mix and execution rules but the person running the test can use any application source code that meets the specifications. Uses exactly the same suite of applications, including the Apache2 web server and the MySQL relational database management system for all tests.
Transaction Types
     Processes inventory transactions. Processes credit card transactions.
Number of Transactions per Screen/Batch
     The computer that drives a TPC test creates batches of transactions that are then sent to the server being tested. Each one of up to 500 web clients submits a single transaction as an HTML request. When that transaction has been processed the client submits another transaction as another HTML request.
Number of Active "Users"
     A formula that varies the number of users based on previous test results. Web client applications submit the transactions to Apache2. The count of active clients ranges from 100-500 during the test sequence.
Reporting of Test Results
     The only number reported is peak transaction throughput. Reports throughput at nine different active user levels. These nine levels show the throughput for cached disk I/O, the throughput for physical disk I/O and the point of transition where cache is exhaused and physical I/O begins to dominate.

Copyright © 2006-2008
Neal Nelson & Associates
Trademarks which may be mentioned on this site are the property of their owners.