Benchmark | Tuesday, September 27, 2011

SPARC T4 Servers Set World Record on Siebel Loyalty Batch

By: Brian Whitney | Principal Software Engineer

Oracle's SPARC T4-2 and SPARC T4-4 servers running Oracle's Siebel Loyalty Batch engine delivered a world record result for batch processing.

  • The SPARC T4-2 and SPARC T4-4 servers running Siebel Loyalty Batch engine, part of Siebel Loyalty Solution, with Oracle Database 11g Release 2 running on Oracle Solaris 10 achieved 7.65M TPH on Accrual (Reward) processing using three Siebel Servers.

  • The world record result was achieved with 24M members and 50M records in the base transaction table.

  • Siebel Loyalty Application was configured with 50 Active Promotions with three Assign Points and four Update Attributes.

  • Oracle's Siebel Server scaled near linearly on SPARC T4 systems achieving 2.72M TPH on a single Siebel Server to 7.65M TPH with three Siebel Servers.

  • The average CPU utilization on the database tier server was 25% and on the application tier server was 65%, leaving significant room for application growth.

Performance Landscape

System Processor TPH Version
3 x SPARC T4-2 (app)
1 x SPARC T4-4 (db)
SPARC T4, 2.85 GHz
SPARC T4, 3.0 GHz
2 x SPARC T3-2 (app)
1 x SPARC T3-1 (app)
1 x SPARC M5000 (db)
SPARC T3, 1.65 GHz
SPARC T3, 1.65 GHz
SPARC64 VII, 2.52 GHz
Customer (app)
Customer (db)
4 x Intel E5540, 2.53 GHz
1 x Itanium, 1.6 GHz
1.5M 8.1.x

Configuration Summary

Hardware Configuration:

3 x SPARC T4-2 servers, each with
2 x SPARC T4 processors, 2.85 GHz
128 GB main memory

1 x SPARC T4-4 server with
4 x SPARC T4 processors, 3.0 GHz
256 GB main memory

1 x Sun Storage 6180 array
16 disk drives
CSM200 with 16 disk drives

Software Configuration:

Oracle Solaris 10
Siebel Server
Oracle Database 11g Release 2 Enterprise Edition

Benchmark Description

Siebel Loyalty enables companies to simulate and process loyalty rewards for their activities across channels and process very high volume accrual and tier assessment transactions via batch process.

The benchmark simulates a workload of Accrual Batch Transactions Processing which imports data through Enterprise Integration Manager (EIM), evaluates eligible promotion and calculates rewards. The key performance metric is transactions per hour (TPH). Key aspects of the workload simulation include:

  • Batch Engine evaluating all accrual promotions and applying all actions in one go,
  • Users do not have control over the sequence in which promotion applied,
  • Promotion actions (assign/redeem points) are rolled back in case of failure.

The number of active promotions and, in particular, the Assign Point action has very significant impact on performance. The load simulated 50 Active promotions with 3 for Assign Points and 7 Update attribute actions configured.

The number of members and the number of queued transactions in the backend database have significant impact on the performance. The benchmark had 24 million members and 52 million records in the base transaction table. The simplified process flow of the benchmark is:

  • calculate accruals base on promotions,
  • credit points to members,
  • initiate any other actions specified in promotions.

See Also

Disclosure Statement

Copyright 2011, Oracle and/or its affiliates. All rights reserved. Oracle and Java are registered trademarks of Oracle and/or its affiliates. Other names may be trademarks of their respective owners. Results as of 9/26/2011.

Join the discussion

Comments ( 0 )
Please enter your name.Please provide a valid email address.Please enter a comment.CAPTCHA challenge response provided was incorrect. Please try again.Captcha

Visit the Oracle Blog


Contact Us


Integrated Cloud Applications & Platform Services