Planning RIS Server Performance

To plan for the needs of your RIS server, you can assess RIS server performance in your test environment prior to deploying RIS in your network. The results you obtain provide some initial performance indications that can assist you in deciding where to make necessary improvements to the RIS serv 626o1411g er configuration.

To determine if you need improvements to enhance RIS server performance, you might first analyze the areas in which potential performance bottlenecks can occur. Doing this in your RIS test environment is beneficial for a baseline performance assessment, because the environment uses multiple server roles on the computer hosting RIS. However, for more definitive assessments, you can monitor performance in your actual environment during times of high demand on your RIS servers.

For more information about the RIS test environment, see "Creating a RIS Test Environment" later in this chapter.

Because RIS primarily uses a file-copy process, RIS server performance factors are common to those of other file input/output (I/O) intensive servers, such as Web, file, and print servers. These performance factors include server throughput, disk throughput, and network speed. When a RIS server or its network connection overloads, the result is increased installation time on client computers and TFTP time-outs during the initial file-copy phase.

For this part of your planning process, use job aid "Planning for RIS Servers" (ACIRIS_02.doc) on the Windows Server 2003 Deployment Kit companion CD (or see "Planning for RIS Servers" on the Web at https://www.microsoft.com/reskit) to indicate whether you plan to use the test or production environment to obtain RIS server performance indications.

Performance Bottlenecks

A bottleneck is the part of a computer system that restricts workflow. Generally, a bottleneck is caused by over-consumption of a specific resource. Some causes of bottlenecks include:

Too many active processes that need access to RAM.

A processor running at a high percentage of utilization.

A disk controller or drive that is slow when accessing data.

To monitor for bottlenecks, you can use the Performance tool provided with the Windows Server 2003 operating system to determine which system components are having an adverse effect on total I/O throughput. You can locate the Performance tool on a Windows Server 2003 from the Start menu in the Administrative Tools group. When you start the tool in MMC, a Help menu item is available to assist you in using the tool.

The components you should consider monitoring include the following:

The best indicator of a memory bottleneck is a sustained high rate of hard page faults. Hard page faults occur when the data a program needs is not found in the working memory visible to the program or elsewhere in physical memory, and must therefore be retrieved from the disk. An acceptable range for this parameter is 0-20 hard page faults per second.

Processor activity is especially important for server-based applications. Two of the most common causes of CPU bottlenecks are CPU-bound applications and excessive interrupts generated by inadequate disk or network subsystem components. Consider processor use in excess of 75 percent to be a bottleneck.

Just as processor usage is the bottleneck for server applications, the disk subsystem is often the bottleneck for file I/O performance. Because a RIS server is file I/O intensive, disk activity is a primary area to analyze. With the Performance tool, you can evaluate the disk activity of your RIS server in terms of total I/O throughput of the server and percentage use of the disk subsystem.

Table  . describes the Performance tool objects and counters you can use to assess performance of critical systems, including memory, processor, and the disk subsystem.

Table  .    Performance Monitor Objects and Counters

Based on the performance monitoring results you obtain, consider making improvements in one or more of the following areas to optimize performance of your RIS servers:

Increase processor speed.

Increase the data transmission rate on the network.

Decrease processor utilization, which can include shutting down unnecessary services on the computer hosting RIS.

Improve disk utilization.

Consider providing a disk configuration that uses multiple SCSI II (minimum) hard drives with Redundant Array of Independent Disks (RAID) 5 disk striping with parity.

Increase the amount of installed RAM.

For this part of your planning process, use job aid "Planning for RIS Servers" (ACIRIS_02.doc) on the Windows Server 2003 Deployment Kit companion CD (or see "Planning for RIS Servers" on the Web at https://www.microsoft.com/reskit) to indicate the components you plan to test (memory, processor, disk subsystem), the potential upgrades you anticipate, and the personnel you want to assign to these tasks.