IBM TEMENOS T24 Workload Optimization PureFlex System

IBM and TEMENOS T24 workload optimization on the new IBM PureFlex System TEMENOS T24 Enterprise performance testing using IBM POWER7 and x86 hybrid architectures with IBM DB2 v 9.7

Auth Author orss Temenos UK Simon Henman & IBM IBM Syste ystems ms and Techno chnolo logy gy Grou Group p ISV Ena Enable blemen ment Mura Murali li Ner Nerall alla Jo John Nutt Nuttle le May May 2012 2012

© Copyright IBM Corporation, 2012

This page intentionally left blank

IBM and Temenos T24 workload optimization on the new IBM PureFlex System

Table of contents Abstract........................................................................................................................................1 Introduction .................................................................................................................................1 Results (monthly close of business) at a glance ..................................................................................... 1

Test environment topology ........................................................................................................2 Physical component layout ......................................................................................................................2 SAN Storage ............................................................................................................................................ 4 Storwize V7000 configuration.................................................................................................................. 5 Easy Tier .................................................................................................................................................. 5 Test methodology .................................................................................................................................... 6 Test description........................................................................................................................................ 6 Data preparation ...................................................................................................................................... 6 Data build process............................................................................................................. 6 Bulk Test .................................................................................................................................................. 7 Bulk Test configuration and setup ..................................................................................... 7 Online test ................................................................................................................................................ 9

Running the tests......................................................................................................................10 Results (COB)........................................................................................................................................ 11 Easy Tier results with TEMENOS T24................................................................................................... 11

Summary....................................................................................................................................13 Resources..................................................................................................................................14 Trademarks and special notices..............................................................................................15


Abstract This white paper describes the testing and optimization of the TEMENOS T24 core banking solution running on the IBM PureFlex System. Temenos, as a long-standing IBM Business Partner IBM, has regularly benchmarked its products on IBM hardware and software as part of that teaming. This paper describes a benchmark of the Temenos T24 product on the IBM PureFlex System. The intended audience for this paper includes technical consultants working with banks on the hardware options available to run a T24 solution and bank employees planning the deployment of T24 on IBM PureFlex.

Introduction TEMENOS T24 is one of the most widely-used international banking systems in the world. It is a 24x7 real-time banking application that provides multiple application server support to a large numbers of banks. As one of the world’s most widely used banking systems, T24 is an open and highly configurable system, incorporating industry-standard practices based on the Temenos implementations in 125 countries. An IBM PureFlex™ System was made available to Temenos for a short period prior to product announcement. A join IBM-Temenos team carried out testing to meet the following objectives:  



 

Investigate the performance of T24E for Java™ and T24 on the new IBM PureFlex platform Test the performance of T24E for Java with IBM WebSphere® Application Server and IBM DB2® on IBM AIX® and Linux® at transaction rates typical of a midsized bank with 5 million accounts. 1 Test the performance of T24E in a hybrid environment with some components of the solution deployed on Intel® nodes and some on IBM POWER7® nodes — the first time T24E has been tested in a hybrid environment Test the performance of T24E for Java with the IBM Storwize® V7000 Easy Tier® feature Compare the performance of T24 and T24E for Java.

Note that hardware requirements and sizing for actual customers are not to be inferred from this paper as other factors, such as architecture, scalability, high availability, additional applications, and growth must also be considered to size a workload correctly.

Results (monthly close of business) at a glance The system processed 5 million accounts capitalization during COB process at the rate of 330 and 345 transactions per second (TPS), respectively (see Table 1). This was also tested using a mix of both architectures.

1

Temenos T24 and Temenos T24E for Java are identical functionally but run in different environments. T24 has its own c-based application server, T24E for Java runs in a JEE application server, such as IBM WebSphere Application Server IBM and Temenos T24 workload optimization on the new IBM PureFlex System

1

Platform

TPS

Threads 36

Application processor 66%

DB processor 30%

POWER7-T24EWebSphere Application Server x86-T24EWebSphere Application Server

330

345

36

60%

40%

Table 1. Results (monthly close of business) at a glance

Test environment topology The integrated design of the PureFlex System starts with a single rack capable of holding up to 14 nodes. Those nodes can be any combination of Intel and IBM POWER® processor-based nodes. In addition, built into the PureFlex System rack, are switches for Ethernet and Fiber Channel (FC) connectivity. The PureFlex System is designed for full redundancy, which allows for multiple switches, fans, and power supplies to ensure that there is no single point of failure. Multiple PureFlex System racks can be combined to satisfy growth and larger configuration requirements.

Physical component layout The PureFlex System used for the Temenos T24 testing used one IBM POWER7 compute node, two Intel compute nodes, and an IBM Storwize V7000 Unified storage system. The PureFlex System chassis management module provides the management capabilities for all the elements in the chassis. The Storwize V7000 Unified storage subsystem is attached through an integrated FC QLogic 8 Gb switch. Figure 1 shows all the components and the ph ysical layout used during this testing.


2

PureFlex System Chassis

Workload drivers

1 0 G b E t h e r n e t s w i t c h

Intel node

Intel node

Power node

8 G b F i b r e C h a n n e l s w i t c h

S t o r w i z 2x8 Gb e V Fibre 7 Channel 0 0 0

PureFlex System chassis management module

Figure 1: The test environment displaying the components and layout

This section provides the software a nd hardware details of each individual partition. Database server partition specifications (POWER7 node):      

POWER7 3.2 GHz 7 cores, 96 GB IBM AIX 7.1 TL01 SP01 One 10 Gb Ethernet (Virtual), Two 8 Gb Fiber Channel (Virtual) DB2 V9.7 fp5 and additional fixes (build 28524) One IBM Storwize V7000

Application server partition specifications (POWER7 node):     

POWER7 3.2 GHz 24 cores, 24 GB AIX 7.1 TL01 SP01 One 10 Gb Ethernet (Virtual) Two 8 Gb Fiber Channel (Virtual) IBM WebSphere Application Server v8 FP2

Virtual I/O server partition specifications (POWER7 node):  

POWER7 3.2 GHz 1 core, 3 GB VIOS 2.2.1.1 IBM and Temenos T24 workload optimization on the new IBM PureFlex System

3

I B M

 

One 10 Gb Ethernet Two 8 Gb Fiber Channel

Application server node specifications (Intel node):    

Intel® 2.7GHz E5-268016 cores, 48 GB RHEL 6.2 64-bit One 10 Gb Ethernet IBM WebSphere Application Server v8 FP2

MQ Server node specifications (Intel node):    

Intel 2.7GHz E5-2680 8 cores, 24 GB RHEL 6.2 64-bit One 10 Gb Ethernet IBM WebSphere MQ v7.1

IBM Storwize V7000 specifications:    

Software version 6.3.0.1 One control enclosure Twenty-two 600 GB 10K rpm SAS hard disk drives (HDDs) Two 300 GB solid-state drives (SSDs)

SAN storage IBM Storwize V7000 Unified is a virtualized storage system designed to consolidate block and file workloads into a single storage system for simplicity of management, reduced cost, highly scalable capacity, performance, and high availability. IBM Storwize V7000 Unified storage also offers improved efficiency and flexibility through built-in SSD optimization, thin provisioning, and non-disruptive migration of data from existing storage. The system can virtualize and reuse existing disk systems, offering a greater potential of return on investment (ROI). Highlights of the Storwize V7000 Unified storage include: 











File support: Consolidates block and file storage in a single system for simplicity and greater efficiency IBM Active Cloud Engine™: Enables highly-efficient policy-based management of files to reduce costs through use of tiered storage Clustered systems: Ability for block systems to both scale up and scale out for performance and capacity. Automated storage tier: Advanced technology for automatically migrating data between storage tiers based on real-time usage analysis patterns New-generation GUI: Easy-to-use data management designed with a graphical user interface (GUI) and point-and-click system management capabilities Virtualized storage: Built-in efficiency technologies virtualization, thin provisioning, Easy Tier, and more


4

Storwize V7000 configuration The internal disk used for the testing resides in 1 c ontrol enclosure with twenty-two 600 GB 10K rpm SAS and two 300 GB SSDs. Storwize V7000 is configured with three 6+P RAID5 managed disks (MDisks) for SAS with stripe size set to 128 and one RAID1 pair for SSDs. These are in a single Easy Tier enabled pool (refer to Figure 2).

Figure 2: Storwize V7000 configuration

Easy Tier Easy Tier is a no-charge feature of the IBM Storwize V7000 system that brings enterprise storage functions to the midrange segment. It enables automated sub-volume data placement throughout different storage tiers to intelligently align the system with current workload requirements and to optimize the usage of SSDs. This functionality includes the ability to automatically and nondisruptively relocate data (at the extent level) from one tier to another tier in either direction to achieve the best available storage performance for your workload in your environment. Easy Tier reduces the I/O latency for hot spots, but it does not replace storage cache. Both Easy Tier and storage cache solves a similar access latency workload problem, but these two methods weigh differently in the algorithmic construction based on locality of reference, recency, and frequency. As Easy Tier monitors I/O performance from the device end (after cache), it can pick up the performance issues that cache cannot solve and complement the overall storage system performance.


5

In general, the I/O of the storage environment is monitored on volumes and the entire volume is always placed inside an appropriate storage tier. Determining the amount of I/O is too complex for monitoring I/O statistics on a single extent and moving them manually to an appropriate storage tier and reacting to workload changes is very tedious task.

Test methodology The primary objective of the test was to measure the performance of T24 Enterprise for Java R12 on IBM DB2 V9.7 in a realistic, typical retail banking environment. The following section describes the T24 test environment and the test cases used.

Test description The transaction test cases and mix percentages were compiled from existing Temenos clients to simulate real-world scenarios and the daily processing load of a typical retail bank with a suitable amount of transaction history. Tests of both online and bulk usage are reported here. The bulk test was of interest accrual and cap italization, or close-of-business (COB) or end-of-day, and consisted of a full T24 month-end close of business run, as per standard T24 benchmarking conditions. The online test was of nine different transaction types (including withdrawals, ATM transactions, and deposits) to represent typical activity during business hours in a retail bank.

Data preparation The data volumes used in the tests are shown in Table 2.

Data set

5 million CAP

Customers

5,000,000

Accounts – all local currency

5,000,000

Branches Transaction history

2000 1 x month full history

Table 2: Data volumes

The data for the benchmark was prepared using a methodology adapted for mid-tier benchmarking. Two sets of T24 R12 model bank were prepared on the JBASE database, 1 million accounts and 5 million accounts, based on the actual account volumes required and these were then imported using the T24 DBImport utility into the IBM DB2 database. Two sets of backup were taken after the database import process. The two sets of data were used to test the scalability of the environment.

Data build process The process involves the following steps: IBM and Temenos T24 workload optimization on the new IBM PureFlex System

6

1. Create scripts for the three transaction types: Customer, Account, and Balances. 2. Run these three scripts in order into T24. 3. Prepare T24 for Accruals and Capitalization. 4. Run COB Accruals till the month-end date. 5. Backup the JBase database in this pre-CAP state. 6. Create a new database in DB2. 7. Set up T24E for Java and run the DBImport utility to convert the JBase data into IBM DB2 data. 8. After the import process, back up the DB2 database. This data build process was done twice for the two sets of account volumes.

Bulk Test The close of business (COB) and capitalization (CAP) process is run at month-end capitalization in which the interest accrued over the month is added to the accounts. 2

This is a bulk process and not the only job that occurs during the month end. This job is typically the most demanding and so it is routinely used to size T24 installations. This workload is typically I/O bound in the database server and processor bound in the application server.

A typical CAP transaction contains over 25 I/Os comprising: 20 selects, 3 inserts and 5 updates, on multiple tables.

Bulk Test configuration and setup The Bulk Test was run as follows: 1. 2. 3.

Two CAP runs were run individually on two environments and the results collected. Both the runs were run from T24E for Java, inside WebSphere Application Server. The T24 on AIX test was run from the AIX environment with 36 threads or agents set up in T24. jDLS (jBASE Distributed Locking) was used for this test. 4. T24E for Java memory locking was set up for the WebSphere Application Server-based) tests. 5. Threads were set to 36+ at the following level for the WebSphere Application Server-based tests. a. T24 workload profile for COB b. Data source connecting to DB2 c. SIBJMSRAThreadPool inside IBM WebSphere Application Server 6. The same database prepared for T24 tests was used for the T24E for Java tests, in order to get a direct comparison of performance of T24 and T24E for Java.

2

We use the term “bulk” rather than “batch” as, in T24, bulk processes are run on the online system while the online system is processing normal work. IBM and Temenos T24 workload optimization on the new IBM PureFlex System

7

7.

Database differences between T24 and T24E: a. T24 used table types of XML and BLOB. b. T24E used table types of XML and CLOB. The BLOB tables of T24 were exported and imported as CLOB types.


8

Online test The set of transaction types for the online sizing activity is listed in Table 3. Channel type

Representative transaction name

T24 transaction

Balance inquiry

Custom routine OFS.GET.ACCT.BA LANCE

Cash withdrawal

OFS CLEARING MANAGER with transaction request

Teller cash deposit

TELLER,MB.DEPOS IT

Outward clearing transactions (collections)

FT,MB.OT

ATM and POS

Branch office transactions

SB account opening (nonmonetary)

ACCOUNT,MB.NEW

OFS-MQ

XML-browser web OFS-MQ XML-browser web OFS-MQ

Balance inquiry Funds transfer (account to account Transfer)

FT,MB.ACCT.TO.AC CT

XML-browser web

Balance inquiry Bank statement Financial transaction Enquiry Clearing

OFS-MQ

Custom Routine OFS.GET.ACCT.BA LANCE

IVR

Internet banking

Trans type interface

Clearing transactions

ENQ ACCT.BAL.TODAY ENQ STMT.ENT.BOOK ENQ STMT.ENT.BOOK

OFS-MQ

OFS CLEARING MANAGER (BULK)

OFS-MQ

OFS-MQ OFS-MQ

Table 3: Transaction set

The online test runs a mixture of these transactions, representing a typical online load. The transactions are injected into T24 using WebSphere MQ. This is a reasonable approach as T24 runs transactions asynchronously even when they are interactive transactions from user interfaces. Please note that the online transactions were run in isolation for proof of concept and collection of metrics for sizing. No formal results (all running at the same time) were taken for measurement during this testing.


9

Running the tests This section describes how the bulk and online tests were run and reports the results of those tests. The following tests were run:     



Running COB through POWER7-WebSphere Application Server Running COB through x86-WebSphere Application Server Running COB through POWER7 and x86- WebSphere Application Server Running COB through TAFC UPM-DB (Capitalization simulation without T24 using the Temenos tool UPM running in Java ) on DB2-JVM Running single-thread tests online (as detailed above on page 8). These tests were streaming thousands of online transactions individually to get the metrics of a single transaction. a. POWER7-WebSphere Application Server b. x86-WebSphere Application Server

The above testing was conducted to prove compatibility and performance profiling. Number 1 of the tests was run to validate the install of T24 Enterprise and to validate the database tables were present and correct. Though all these tests were run, the results of prime importance were those of running COB CAP with POWER7, x86 on 5 million accounts and an online test of POWER7 and x86 on 5 million accounts.


10

Results (COB) The system processed 5 million accounts capitalization during COB process at the rate of 330 and 345 transactions per second (TPS) and in excess of 500 TPS in a mixed-architecture workload. These transaction rates would meet the requirements of a typical 5 million account retail bank, and clearly the PureFlex System as a whole is capable of supporting much larger workloads (see Table 4).

Platform

TPS

Threads 36

Application processor 66%

DB processor 30%

POWER7-T24EWebSphere Application Server x86-T24EWebSphere Application Server POWER7-TAFC

330

345

36

60%

40%

402

36

54%

55%

Table 4. Summary of results

The results show that the T24 Enterprise application can run on WebSphere Application Server and share the architecture between different server types, ea ch running WebSphere Application Server, to enable it to process workload to a single database instance running in Power 7. The flexibility this shows enables T24 Enterprise to run as a single application server or multiple application servers sharing database connections through WebSphere Application Server. The POWER7-TAFC result is only used as a basic comparison. TAFC is the traditional way T24 is executed on several architectures. This was used to refine the table types in DB2 that should be used in T24 Enterprise for performance testing. The above results show the new Temenos offering of T24 Enterprise running achieved approximately 20% lower throughput in these tests than the traditional workload processing of TAFC. The Java workload (T24 Enterprise) has been shown in other tests to exceed the performance of the TAFC implementation and the discrepancy may be because in these tests there was only time for very basic tuning. Due to the way the memory was allocated between server types, x86 and POWER 7, you can see that the x86 server running less cores was able to o utperform the POWER 7 application server. The team believes this is due to the spread of memory between the nodes.

Easy Tier results with TEMENOS T24 Initial testing prior to enabling Easy Tier showed spikes in response times. These spikes peaked at over 400 ms. Easy Tier was enabled and resulted in reductions in response times and improved throughput. The combination of HDD and SSD with Easy Tier activated resulted in the best drive utilization, response time, and processor (user, sys, wait) times. The test team also observed that there is no nee d to manually tune the storage to see the performance benefits; Easy Tier takes care of it automatically. The test team had also observed the ease with which Easy Tier can be configured and deployed.


11

With the IBM Storage Tier Advisor Tool and Easy Tier, the users are now able to provision SSDs accurately for their workloads. The Storage Tier Advisor Tool provides a high-level summary of workload characteristics and hot extents of the volumes that are monitored. It also provides assistance for SSD capacity planning with Easy Tier. When Easy Tier is enabled, it observes the I/O pattern, and then migrates the data from the HDD MDisks to the SSD MDisks non-disruptively. This combination of Easy Tier and the Storage Tier Advisor Tool provides a high impact, combined with low investment of efforts for the user (refer to Figure 3 and Figure 4).

Figure 3. Easy Tier results

Figure 4. Easy Tier results (continued)


12

Summary After careful consideration of the performance profile results: 







This testing validated IBM PureFlex System as a platform for both T24 and T24E for Java. Performance testing showed the IBM PureFlex System performing exceptionally well and easily capable of handling the workloads of typical T24 customers. The sizing data should be considered preliminary data only. This testing was run on prerelease hardware and firmware and there was very limited time available for performance optimization. EasyTier storage optimization and the V7000 storage subsystem provide a big improvement in disk response times, which can translate into faster transaction response times and faster batch runtimes.


13

Resources These websites provide useful references to supplement the information contained in this paper: 







IBM Systems on PartnerWorld® ibm.com/partnerworld/systems Virtual Loaner Program ibm.com/systems/vlp IBM Redbooks® ibm.com/redbooks Temenos T24 and T24 Enterprise www.temenos.com


14

Trademarks and special notices © Copyright IBM Corporation 2012. References in this document to IBM products or services do not imply that IBM intends to make them available in every country. IBM, the IBM logo, and ibm.com are trademarks or registered trademarks of International Business Machines Corporation in the United States, other countries, or both. If these and other IBM trademarked terms are marked on their first occurrence in this information with a trademark symbol (® or ™), these symbols indicate U.S. registered or common law trademarks owned by IBM at the time this information was published. Such trademarks may also be registered or common law trademarks in other countries. A current list of IBM trademarks is available on the Web at "Copyright and trademark information" at www.ibm.com/legal/copytrade.shtml. Java and all Java-based trademarks and logos are trademarks or registered trademarks of Oracle and/or its affiliates. Microsoft, Windows, Windows NT, and the Windows logo are trademarks of Microsoft Corporation in the United States, other countries, or both. Intel, Intel Inside (logos), MMX, and Pentium are trademarks of Intel Corporation in the United States, other countries, or both. UNIX is a registered trademark of The Open Group in the United States and other countries. Linux is a trademark of Linus Torvalds in the United States, other countries, or both. Other company, product, or service names may be trademarks or service marks of others. Information is provided "AS IS" without warranty of any kind. All customer examples described are presented as illustrations of how those customers have used IBM products and the results they may have achieved. Actual environmental costs and performance characteristics may vary by customer. Information concerning non-IBM products was obtained from a supplier of these products, published announcement material, or other publicly available sources and does not constitute an endorsement of such products by IBM. Sources for non-IBM list prices and performance numbers are taken from publicly available information, including vendor announcements and vendor worldwide homepages. IBM has not tested these products and cannot confirm the accuracy of performance, capability, or any other claims related to non-IBM products. Questions on the capability of non-IBM products should be addressed to the supplier of those products. All statements regarding IBM future direction and intent are subject to change or withdrawal without notice, and represent goals and objectives only. Contact your local IBM office or IBM authorized reseller for the full text of the specific Statement of Direction. Some information addresses anticipated future capabilities. Such information is not intended as a definitive statement of a commitment to specific levels of performance, function or delivery schedules with respect to any future products. Such commitments are only made in IBM product announcements. The information is


15

presented here to communicate IBM's current investment and development activities as a good faith effort to help with our customers' future planning. Performance is based on measurements and projections using standard IBM benchmarks in a controlled environment. The actual throughput or performance that any user will experience will vary depending upon considerations such as the amount of multiprogramming in the user's job stream, the I/O configuration, the storage configuration, and the workload processed. Therefore, no assurance can be given that an individual user will achieve throughput or performance improvements equivalent to the ratios stated here. Photographs shown are of engineering prototypes. Changes may be incorporated in production models. Any references in this information to non-IBM websites are provided for convenience only and do not in any manner serve as an endorsement of those websites. The materials at those websites are not part of the materials for this IBM product and use of those websites is at your own risk.


16

IBM TEMENOS T24 Workload Optimization PureFlex System

Recommend Documents