Optimizing Oracle's Siebel Applications for the Sun Platform, Solaris 10 OS, and UltraSPARC T1 Processor

Print-friendly Version

Abstract

This paper discusses the performance optimization of a complete Siebel Enterprise solution on the Sun platform using UltraSPARC T1 and UltraSPARC IV+ processor based servers. Also presented in detail is information on tuning the Solaris Operating System (OS), Siebel 7.7 software, Oracle Database 10g server, Sun storage products, and Sun Java System Web Server. In addition, the article discusses unique features of the Solaris 10 OS that reduce risk while helping improve the performance and stability of Siebel applications.

For more details, see the list of contents below.

Download the document as PDF.

Contents

• Introduction
• Summary
• Oracle's Siebel Application Architecture Overview
• Optimal Architecture for Benchmark Workload
• Hardware and Software Used
• Workload Description
• OLTP Components (Siebel Web Thin-Client End Users)
• Batch Server Components
• 8,000 Concurrent Users Test Results Summary
• Response Times and Transaction Throughput
• Server Resource Utilization
• 12,500 Concurrent Users Test Results Summary
• Response Times and Transaction Throughput
• Server Resource Utilization
• Scalability on the Sun Platform
• Performance Tuning
• Tuning Solaris OS for Siebel Server
• Solaris Tuning for Siebel using libumem
• Solaris MPSS (Multiple Page Size Support) Tuning for Siebel
• The Solaris Kernel and TCP/IP Tuning Parameters for Siebel Server
• Tuning Siebel Server for Solaris
• Tuning FINS Call Center and eChannel Siebel Modules
• Siebel EAI-HTTP Adapter
• Making Siebel Applications Use the Maximum Server Capacity
• The Siebel MaxTasks Upper Limit Debunked
• Why Are End Users Complaining About Lost Connections?
• Tuning Sun Java System Web Server
• Tuning the SWSE (Siebel Web Server Extension)
• Tuning Siebel Standard Oracle Database and Sun Storage
• Optimal Database Configuration
• Properly Locating Data on the Disk for Best Performance
• Disk Layout and Oracle Data Partitioning
• Solaris MPSS Tuning for the Oracle Database Server
• Hot Table Tuning and Data Growth
• Oracle Parameters Tuning
• Solaris Kernel Parameters on Oracle Database Server
• SQL Query Tuning
• Rollback Segment Tuning
• Database Connectivity Using Host Names Adapter
• High I/O With Oracle Shadow Processes Connected to Siebel
• Siebel Database Connection Pooling
• Performance Tweaks That Provide No Gains
• Tips and Scripts for Diagnosing Oracle's Siebel on the Sun Platform
• Monitoring Siebel Open Session Statistics
• Listing the Parameter Settings for a Siebel Server
• Finding All OMs Currently Running for a Component
• Finding the Number of Active Servers for a Component
• Finding the Tasks for a Component
• Setting Detailed Trace Levels on the Siebel Server Processes
• Finding the Number of GUEST Logins for a Component
• Calculating the Memory Usage for an OM
• Finding the Log File Associated With a Specific OM
• Producing a Stack Trace for the Current Thread of an OM
• Showing System-Wide Lock Contention Issues Using lockstat
• Showing the Lock Statistic of an OM Using plockstat
• How To "Truss" an OM
• How to Trace the SQL Statements for a Siebel Transaction
• Changing the Database Connect String
• Enabling and Disabling Siebel Application Components
• Appendix A: Transaction Response Times
• Appendix B: Database Objects Growth During the Test
• References
• Acknowledgements
• About the Author