MIIS 2003 Capacity Planning Test Summary - Memory

Applies To: Windows Server 2003 with SP1

Previous Sections in This Guide

This section discusses tests performed to determine how MIIS 2003 utilizes memory and presents the results.

Test Description

Staging and synchronization operations for groups of objects were tested. The tests used a management agent to project groups of objects into the metaverse. These groups of objects were tested in various increments, from 10,000 objects to 500,000 objects. Nine additional management agents then performed a number of join operations with those objects based on indexed metaverse attributes. Memory was monitored during these tests to determine whether or not any predictable relationship existed between the number of objects being processed and the memory used.

Expected Results

Microsoft SQL Server 2000 limits memory usage to 2 GB by default. It is expected that the server will utilize the effective default of 2 GB RAM and this limit will be reached faster when processing a larger number of objects.

Test Results Summary

The tests did confirm that the 2 GB addressable limit was reached by Microsoft SQL Server 2000. As expected, the limit was reached faster when larger batches of objects were processed.

At a minimum, servers should be configured with at least 1 GB RAM for each processor. If you will be managing 50,000 or more objects from more than two data sources, then it is recommended that you use 2 GB RAM for each processor.

Test Scenario

The test environment consisted of a single server configured with ten management agents and run profiles configured to perform staging and synchronization operations for groups of user objects that ranged in size from 10,000 to 500,000 objects. The details of the server configuration and test results are described below.

Server Hardware Configuration

A single server configuration was used to perform the tests. Both MIIS 2003 and Microsoft SQL Server 2000 were installed on the same server. For the purposes of this discussion, the server configuration will be referred to as Q30. As in previous tests, Q30 indicates the number of processors (Q for a quad-processor server) followed by two digits to indicate clock speed. Details of the server platform are summarized in the table below.

Table 23: Server hardware configuration

Server Designation Model Configuration

Q30

HP DL580 G2

Quad Intel Xeon MP @ 3.0 GHz

Hyper-threading

4 GB System RAM

Internal 5i SCSI controller

4x 36.4 GB Ultra 320 HDD (10k)

Windows Server 2003, Enterprise Edition

Advanced Windowing Extensions (AWE) not enabled

MIIS 2003 Configuration

MIIS 2003 with Service Pack 1 was installed on the server being tested. An increasing number of management agents were executed on the test platform to explore the memory usage of MIIS 2003 and SQL server under increasing load.

Database Configuration

On the Q30 test platform, Microsoft SQL Server 2000 with Service Pack 3a was installed with the SQL database and SQL log files residing on the D: drive and E: drive respectively. These volumes were created on a single 36.4 GB Ultra 320 hard disk.

Management Agent Configuration

Ten management agents were installed on MIIS 2003. The following table summarizes the configuration of each one.

Table 24: Management agent configuration

Management Agent Name Type Notes

TXT01

Text File

  • Local data source (text files used as the data source are located on the MIIS 2003 server being tested to avoid impact by network latency)

  • Delimited text file

  • Used to project objects into the metaverse.

  • Variable range of objects (10,000 – 50,000 - 100,000 – 200,000 and 500,000)

  • Each user object has 25 attributes defined.

  • No multivalued attributes.

TXT02 - TXT10

Text File

  • Used to join objects to the existing metaverse objects.

  • In all other respects, they are configured the same as TXT01.

Run Profile Configuration

For each management agent, two run profiles were created for the staging and synchronization operations: a full import run profile for staging and a full synchronization run profile for projections and joins. As with previous tests, this was done because the MIIS 2003 metaverse was reset at the beginning of each test. These full import and synchronization profiles were necessary to populate the database. During the testing process, all run profiles were executed sequentially. No concurrent operations were tested.

Test Results

During the test runs, user objects from the first management agent were first staged and then added to the metaverse. The remaining nine management agents were then run and their objects were joined to those already in the metaverse. Memory utilization was monitored during these processes. The following different batches of user objects were tested: 10,000, 50,000, 100,000, 200,000 and 500,000. The results of the 10,000 object test and the 50,000 object test are shown below.

Chart: Avg committed bytes in useChart: Avg committed bytes in use

For groups of more than 50,000 objects, MIIS 2003 memory utilization exceeded 1 GB RAM but did not exceed the 2-GB RAM limit. Once the ninth management agent was processed, MIIS utilized the 2-GB RAM limit every time.

Observations

  • When groups of 10,000 objects were processed, memory usage did not reach 1 GB even when all ten management agents had been processed.

  • When groups of 50,000 objects were processed, 1 GB was exceeded as soon as the second management agent was run. The 2-GB limit was reached when the tenth management agent was processed.

  • It was found that although the limit of 2GB was reached when groups of 100,000 and 200,000 objects were processed, the number of pages errors remained at zero.

Next

See Also

Concepts

Introduction to MIIS 2003 Capacity Planning
MIIS 2003 Capacity Planning Test Summary - Processor
MIIS 2003 Capacity Planning Test Summary - SQL Server
MIIS 2003 Capacity Planning Test Summary - Disk Performance
MIIS 2003 Capacity Planning Test Summary - Database Size
MIIS 2003 Capacity Planning Test Summary - Network
MIIS 2003 Capacity Planning - Additional Performance Considerations