TEMENOS T24, Microsoft SQL Server 2012, and ... - TechNet Blogs

High-Water Benchmark Testing OverviewThe objective of the high-water benchmark testing was to measure the online business and close ofbusiness (COB) transaction scalability and throughput capacity of a T24 R12 retail Model Bank onWindows Server 2008 R2, SQL Server 2012, and Intel Xeon processor-based servers. Test cases andmix percentage were compiled from existing and prospective T24 customers to simulate the realworldbanking scenarios.The benchmark testing looked at two key requirements for the core system of banks: performanceand scalability.Two types of tests were run to measure T24 performance:Online business scenario test (online test)The online business test measured the capability of T24 to process online businesstransactions to represent typical activity during business hours in a large retail bank. Thetest used nine different types of transactions, including withdrawals, ATM transactions, anddeposits.Interest accrual and capitalization test (COB test)The COB test measured the capability of T24 to run the end-of-day batch processes for 25million accounts. The test consisted of a full T24 month-end close-of-business run, as perstandard T24 benchmarking conditions. The primary processing job measured was theaccrual and capitalization of account interest.Targets for the benchmark testing were set by the Temenos Retail Banking Committee:The COB accrual and capitalization target was set to 7,000 TPS.The online testing target was set to 7,000 TPS during a one-hour period of peak activity ofthe core system. The maximum CPU utilization of database server during peak hour was not to exceed 75percent.Benchmark Testing EnvironmentThe benchmark testing environment was designed to emulate a realistic, large-scale retail bankingenvironment.The tests used Intel Xeon processor-based servers from Hewlett-Packard (HP) for the applicationand database tiers. The tests were run on 25 application servers connected to a central, singleinstance of SQL Server 2012 and Windows Server 2008 R2 Enterprise. The storage hardware usedwas the XIO Hyper ISE 14.4H. Hyper ISE uses Continuous Adaptive Data Placement (CADP)algorithms to identify and accelerate hot-spots on storage caused by the database. By continuouslymonitoring these hot-spots in real time, data is placed in the right storage media at the right time,with no administrative intervention required.Benchmark Testing Results: TEMENOS T24, Microsoft SQL Server 2012, and Intel Xeon Processor-Based Servers 7