r6 - 2014-09-10 - 15:41:35 - Main.gcovellYou are here: TWiki >  Deployment Web > DeploymentPlanningAndDesign > PerformanceDatasheetsAndSizingGuidelines > CollaborativeLifecycleManagementPerformanceReportRDNG501Release

new.png Collaborative Lifecycle Management performance report: RDNG 5.0.1

Authors: GustafSvensson
Date: September 9, 2014
Build basis: Rational Doors Next Generation 5.0.1

Introduction

This report compares the performance of Rational DOORS Next Generation version 5.0.1 to the previous Rational Requirements Composer version 5.0 release. The test objective is achieved in three steps:

  • Run version 5.0 with standard 1.5 hour test using 400 concurrent users.
  • Run version 5.0.1 with standard 1.5 hour test using 400 concurrent users.
  • The test is run three times for each version and the resulting six tests are compared with each other. Three tests per version is used to get a more accurate picture since there are variations expected between runs.

Disclaimer

The information in this document is distributed AS IS. The use of this information or the implementation of any of these techniques is a customer responsibility and depends on the customer’s ability to evaluate and integrate them into the customer’s operational environment. While each item may have been reviewed by IBM for accuracy in a specific situation, there is no guarantee that the same or similar results will be obtained elsewhere. Customers attempting to adapt these techniques to their own environments do so at their own risk. Any pointers in this publication to external Web sites are provided for convenience only and do not in any manner serve as an endorsement of these Web sites. Any performance data contained in this document was determined in a controlled environment, and therefore, the results that may be obtained in other operating environments may vary significantly. Users of this document should verify the applicable data for their specific environment.

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 many factors, including considerations such as the amount of multi-programming 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 results similar to those stated here.

This testing was done as a way to compare and characterize the differences in performance between different versions of the product. The results shown here should thus be looked at as a comparison of the contrasting performance between different versions, and not as an absolute benchmark of performance.

What our tests measure

We use predominantly automated tooling such as Rational Performance Tester (RPT) to simulate a workload normally generated by client software such as web browsers. All response times listed are those measured by our automated tooling and not a client.

The diagram below describes at a very high level which aspects of the entire end-to-end experience (human end-user to server and back again) that our performance tests simulate. The tests described in this article simulate a large part of the end-to-end transaction as indicated. Performance tests include some simulation of browser rendering and network latency between the simulated browser client and the application server stack.

schematicperformancetestmap_somebrowsersim.jpg

Findings

Performance goals

  • Verify that there are no performance regressions between current release and prior release with 400 concurrent users using the workload described below.
    • When comparing the three 5.0.1 runs with the three 5.0 runs the average of the three runs is used for comparison. Generally a 10% difference is accepted.

Findings

  • Performance
    • The performance test does not show much change between 5.0.1 and 5.0.

Topology

The topology under test is based on Standard Topology (E1) Enterprise - Distributed / Linux / DB2.

Server Overview

The specifications of machines under test are listed in the table below. Server tuning details listed in Appendix A

Function Number of Machines Machine Type CPU / Machine Total # of CPU Cores/Machine Memory/Machine Disk Disk capacity Network interface OS and Version
Proxy Server (IBM HTTP Server and WebSphere Plugin) 1 IBM System x3250 M4 1 x Intel Xeon E3-1240 3.4GHz (quad-core) 8 16GB RAID 1 -- SAS Disk x 2 279GB Gigabit Ethernet Red Hat Enterprise Linux Server release 6.3 (Santiago)
JTS Server 1 IBM System x3550 M4 2 x Intel Xeon E5-2640 2.5GHz (six-core) 24 32GB RAID 5 -- SAS Disk x 4 279GB Gigabit Ethernet Red Hat Enterprise Linux Server release 6.3 (Santiago)
RDNG Server 1 IBM System x3550 M4 2 x Intel Xeon E5-2640 2.5GHz (six-core) 24 32GB RAID 5 -- SAS Disk x 4 279GB Gigabit Ethernet Red Hat Enterprise Linux Server release 6.3 (Santiago)
Database Server 1 IBM System x3650 M4 2 x Intel Xeon E5-2640 2.5GHz (six-core) 24 64GB RAID 10 -- SAS Disk x 16 279GB Gigabit Ethernet Red Hat Enterprise Linux Server release 6.3 (Santiago)
RPT Workbench 1 VM image 2 x Intel Xeon X7550 CPU (1-Core 2.0GHz 64-bit) 2 6GB SCSI 80GB Gigabit Ethernet Microsoft Windows Server 2003 R2 Standard Edition SP2
RPT Agent 1 xSeries 345 4 x Intel Xeon X3480 CPU (1-Core 3.20GHz 32-bit) 4 3GB SCSI 70GB Gigabit Ethernet Microsoft Windows Server 2003 Enterprise Edition SP2
RPT Agent 1 xSeries 345 4 x Intel Xeon X3480 CPU (1-Core 3.20GHz 32-bit) 4 3GB RAID 1 - SCSI Disk x 2 70GB Gigabit Ethernet Microsoft Windows Server 2003 Enterprise Edition SP2
RPT Agent 1 Lenovo 9196A49 1 x Intel Xeon E6750 CPU (2-Core 2.66GHz 32-bit) 2 2GB SATA 230GB Gigabit Ethernet Microsoft Windows Server 2003 Enterprise Edition SP2
Network switches N/A Cisco 2960G-24TC-L N/A N/A N/A N/A N/A Gigabit Ethernet 24 Ethernet 10/100/1000 ports

Network connectivity

All server machines and test clients are located on the same subnet. The LAN has 1000 Mbps of maximum bandwidth and less than 0.3ms latency in ping.

Data volume and shape

The artifacts were distributed between 10 projects for a total of 240,659 artifacts.

The repository contained the following data:

  • 219 modules
  • 218,632 module artifacts
  • 22,027 requirement artifacts
  • 11,715 folders
  • 150 collections
  • 704 reviews
  • 114,258 comments
  • 3000 public tags
  • 500 private tags
  • 2,374 terms
  • 75,941 links
  • 2000 views

  • JTS index size on disk = 1.6 GB
  • RM index size on disk = 17.3 GB

The large project contained the following data:

  • 22 modules
  • 1,170 folders
  • 2,171 requirement artifacts
  • 20,600 module artifacts
  • 15 collections
  • 62 reviews
  • 11,350 comments
  • 300 public tags
  • 50 private tags
  • 132 terms
  • 7,676 links

Methodology

Rational Performance Tester was used to simulate the workload created using the web client. Each user completed a random use case from a set of available use cases. A Rational Performance Tester script is created for each use case. The scripts are organized by pages and each page represents a user action.

Based on real customer use, the test scenario provides a ratio of 70% reads and 30% writes. The users completed use cases at a rate of 30 pages per hour per user. Each performance test runs for 90 minutes after all of the users are activated in the system.

Test cases and workload characterization

Use case Description % of Total Workload
Login Connect to the server using server credentials. None
Create a collection Create collections with 10 artifacts. 4
Filter a query Run a query that has100 results and open 3 levels of nested folders. 3
Open nested folders Create review and complete review process. 3
Manage folders Create a folder, move it to a new location, and then delete the folder. 1
Query by ID Search for a specific ID in the repository. 5.5
View collections View collections that contain 100 artifacts from the collections folders. 10
Create a multi-value artifact Create a multi-value artifact and then add a multi-value attribute. 2
Create an artifact Create a requirement that contains a table, an image and rich text. Edit an artifact that has 100 enumerated attributes and modify an attribute. 2
Show artifacts in a Tree view Open a folder that contains artifacts with links and show the data in a tree view. 6
Open graphical artifacts Open business process diagrams, use cases, parts, images, sketches and story boards. 4
Create and edit a storyboard Create and edit a storyboard. 3.5
Display the hover information for a collection Open a collection that contains 100 artifacts and hover over the Artifacts page. 1.5
Query by String Search for a string that returns 30 matched items. 7.5
Create a PDF report Generate a 50-artifact PDF report. 1.5
Create a Microsoft Word report Generate a 100-artifact Microsoft Word report. 1.5
Browse project folder Open module folder and hover over large module 2.5
Open collection and display first page Open filterable data collection page. 1.5
Create artifact in large module Create an artifact from within a module. 1.5
Display module history Open module and display history and audit history. 1
Create traceability report using 50 artifacts Select 50 artifacts and generate report. 1.5
Open large module Open a large module and jump to the bottom. 7
Scroll through a large module Open a large module and scroll down to the bottom. 7
Swich between modules First open medium sized module and then open a large module. 6.5
Edit a module artifact Open a large module and inline edit a child artifact. 4
Save a module Edit a module and save. 1
Create a large module report Open a module and generate a report. 1
Upload 4MB file in new artifact Upload 4MB file. 3.5
Switch view Add columns and filters to view. 1
Hover over linked artifact Open artifact with 50 links and hover over linked artifact. 1

Response time comparison

The median response time provided more even results than the average response time. The nature of the high variance between tests where some tasks at time takes a longer time to run, such as when the server is under heavy load, makes the average response time less predictive. Both the median and average values are included in the following tables and charts for comparison.

In the repository that contained 240,000 artifacts with 400 concurrent users, no obvious regression was shown when comparing response times between runs.

The numbers in the following chart include all of the pages for all of the scripts that ran.

Results

Observation

  • Memory usage is the same in 5.0.1 as in 5.0. Average CPU usage is less in 5.0.1.

Garbage collection

Verbose garbage collection is enabled to create the GC logs. The GC logs show very little variation between runs. Below is an example of the output from the GC logs for JTS and RDNG, including versions 5.0.1 and 5.0 for each application server.

RDNG

  • Observation: The graphs show similar behavior in 5.0.1 as in 5.0.

  • RDNG 5.0
  • RDNG 5.0.1

JTS

  • Observation: The graphs show similar behavior in 5.0.1 as in 5.0

  • JTS 5.0
  • JTS 5.0.1

Create a collection

Back to Test Cases & workload characterization

Filter a query

Back to Test Cases & workload characterization

Open nested folders

Back to Test Cases & workload characterization

Manage folders

Back to Test Cases & workload characterization

Query by ID

Back to Test Cases & workload characterization

View collections

Back to Test Cases & workload characterization

Create a multi-value artifact

Back to Test Cases & workload characterization

Create an artifact

  • Observation: When averaging out the three runs the create an artifact action is about 12% faster in 5.0.1. Saving the artifact is also faster (around 10%)

Back to Test Cases & workload characterization

Show artifacts in a Tree view

Back to Test Cases & workload characterization

Open graphical artifacts

Back to Test Cases & workload characterization

Create and edit a storyboard

Back to Test Cases & workload characterization

Display the hover information for a collection

Back to Test Cases & workload characterization

Query by string

Back to Test Cases & workload characterization

Create a PDF report

Back to Test Cases & workload characterization

Create a Microsoft Word report

Back to Test Cases & workload characterization

Browse project folder

Back to Test Cases & workload characterization

Open collection and display first page

Back to Test Cases & workload characterization

Create artifact in large module

Back to Test Cases & workload characterization

Display module history

Back to Test Cases & workload characterization

Create traceability report using 50 artifact

Back to Test Cases & workload characterization

Open large module

Back to Test Cases & workload characterization

Scroll through a large module

Back to Test Cases & workload characterization

Switch between modules

Back to Test Cases & workload characterization

Edit a module artifact

  • Observation: When averaging out the three runs the save an artifact action is about 12% faster in 5.0.1.

Back to Test Cases & workload characterization

Save a module

Back to Test Cases & workload characterization

Create a large module report

Back to Test Cases & workload characterization

Uploading 4 MB file in new artifact

  • Observation: When averaging out the three runs the action of saving the new artifact is about 12% faster in 5.0.1.

Back to Test Cases & workload characterization

Switch view

Back to Test Cases & workload characterization

Hover over linked artifact

Back to Test Cases & workload characterization

Appendix A

Product
Version Highlights for configurations under test
IBM HTTP Server for WebSphere Application Server 8.5.5.1 IBM HTTP Server functions as a reverse proxy server implemented via Web server plug-in for WebSphere Application Server. Configuration details can be found from the CLM infocenter.

HTTP server (httpd.conf):

Web server plugin-in (plugin-cfg.xml):

  • ServerIOTimeout="900"

OS Configuration:

  • max user processes = unlimited
IBM WebSphere Application Server Network Deployment 8.5.5.1 JVM settings:

  • GC policy and arguments, max and init heap sizes:

-XX:MaxDirectMemorySize=2g -Xgcpolicy:gencon -Xmx8g -Xms8g -Xmn2g -Xcompressedrefs -Xgc:preferredHeapBase=0x100000000 -Xverbosegclog:logs/gc.log

Thread pools:

  • Maximum WebContainer = 200
  • Minimum WebContainer = 200

OS Configuration:

System wide resources for the app server process owner:

  • max user processes = unlimited
  • open files = 65536

DB2 DB2 10.1
LDAP server
License server Hosted locally by JTS server
RPT workbench 8.3 Defaults
RPT agents 8.5.1 Defaults
Network Shared subnet within test lab

For more information

About the authors

GustafSvensson


Questions and comments:
  • What other performance information would you like to see here?
  • Do you have performance scenarios to share?
  • Do you have scenarios that are not addressed in documentation?
  • Where are you having problems in performance?

Warning: Can't find topic Deployment.PerformanceDatasheetReaderComments

Edit | Attach | Printable | Raw View | Backlinks: Web, All Webs | History: r6 < r5 < r4 < r3 < r2 | More topic actions
 
This site is powered by the TWiki collaboration platformCopyright © by IBM and non-IBM contributing authors. All material on this collaboration platform is the property of the contributing authors.
Contributions are governed by our Terms of Use. Please read the following disclaimer.
Dashboards and work items are no longer publicly available, so some links may be invalid. We now provide similar information through other means. Learn more here.