Appendix A: Raymond James Lab Report

Overview

On October 10, 2005, Raymond James, Hewlett-Packard, and Microsoft went to the lab to evaluate the feasibility of upgrading the Raymond James data warehouse to SQL Server 2005. The evaluation took place in the HP labs on a Superdome and SAN environment similar to the one owned by Raymond James. One instance of SQL Server 2000 SP3 and one instance of SQL Server 2005 September CTP were installed and restored on a twelve-1.5GHz processor Superdome with 24GB of RAM and approximately 9TB of disk. The upgrade was performed by attaching and restoring the SQL 2000 SP3 databases to the 2005 instance. The objective of this test was for Raymond James to evaluate the amount of effort required to achieve benefits in areas of performance gains and reduction in maintenance time. The only preparation performed on the environment was that the statistics were updated on the SQL Server 2005 instance. Both of the instances shared the same resources, but only one of the instances at a time was active during the below tests. This ensured that only the versions of SQL Server were being evaluated as part of this test, since they were both running on the same hardware.

Please note that Raymond James and HP tracked resources consumes and time for each test. We obtained permission from Raymond James to share their findings along with their observations and executive summary.

Note

Raymond James Financial (NYSE-RJF) is a Florida-based diversified holding company providing financial services to individuals, corporations, and municipalities through its subsidiary companies. Its three wholly owned broker/dealers, Raymond James & Associates, Raymond James Financial Services, and Raymond James Ltd. have more than 4,800 financial advisors serving 1.2 million accounts in 2,200 locations throughout the United States, Canada, and overseas. In addition, total client assets are approximately $163 billion, of which approximately $30 billion are managed by the firm's asset management subsidiaries.

During the activity the Raymond James team set out and completed the following objectives:

• Migration of existing SQL2000 Data

• Installation Considerations

• Determination of the data transformation tools by comparing SQL Server Integration Services (SSIS) offering in SQL2005 versus Data Transformation Services (DTS)

• Extract Transact and Load (ETL) comparison of the database engines

• Speed of Relational Queries in SQL2005 versus SQL2000

• Database Maintenance timings

Raymond James made the following observations:

• Easy Migration

• Trouble-free Installation

• SSIS: The configuration options are much more intuitive and user-friendly when compared with DTS.

• ETL: Single day runs are equal, but 2005 runs faster with longer multi-day runs. However, there exist obvious differences in how SQL 2005 interprets SQL code within stored procedures when migrating stored procedures "as is." Some run faster as is, and some run longer as is. This indicates certain performance T-SQL enhancements that can probably be utilized to even further improve performance once stored procedures are migrated into 2005.

• Relational Queries: 2005 provided much faster query response times, with an average rate of four times faster without any T-SQL enhancements. Only one stored procedure ran slower, lending more weight to the notion that 2005 T-SQL enhancements can be utilized to further improve our KPI reports.

• Database Maintenance: Overall, no real difference.

Figures A-1 through A-7 show the results of the various tests.

Figure A-1

Figure A-2

Figure A-3

Figure A-4

Figure A-5

Figure A-6

Figure A-7