This section describes how to use the server hardware sizing information you gathered to plan and create a pilot project. You will need to size server hardware within the pilot project and mimic your production environment in a test lab.
At the end of this section, you will be prepared to run your pilot project to collect data and determine your individual hardware system load signatures.
The numbers given for hardware sizing in “Determining Site System Hardware Requirements” earlier in this chapter are for a production environment. When running a pilot project, you must adjust these numbers.
Pilot projects require you to resize site system hardware because you typically cannot conduct tests using hardware that is identical to what you will use in your production environment. You should test with computers that have slightly more memory and disk space than those you will use in a production environment. Doing so will give you an idea of performance limits without quickly exceeding the limits of your hardware. If you choose to test with computers sized for a production environment, you might encounter serious performance issues when you enable enough features and objects to push the hardware limits of your site hierarchy design, thereby causing you to collect incorrect data about your SMS site.
Having slightly more memory and disk space will also help you obtain a representative load signature. Because the SMS log files will be running during your pilot project and not during your actual SMS deployment, their presence can generate a false load signature in production computers. Slightly larger hardware will allow some overhead for the log files (up to 10%), and thus provide a more accurate load signature. For more information about enabling SMS log files, see “Status Messages Versus Logging” in Chapter 16, “Introducing the SMS 2.0 Flowcharts.”
For the test results to be useful, it is important that you create a test environment that truly reflects your intended site hierarchy, with representative network connections intact and available for testing. Use the network diagram you created at the beginning of the project to help you. Add at least one client for each client platform that you plan to support.
To be successful, your pilot project site hierarchy should include a representative of each type of site system role, server, and client that will exist in your full SMS site hierarchy. The central site and several of its child primary sites are obvious choices; however, several secondary sites and clients are also important. Your pilot project data might be distorted if you simply create a test environment that contains only computers with fast processors and a lot of memory and not the less powerful, smaller computers in the hierarchy. You can gain valuable information by watching how your site hierarchy will function in your network as it exists today — including outdated computers and obsolete applications.
The object of the pilot project is to create a small-scale version of your full planned deployment, not to watch SMS function in a clean environment with few obstacles. The closer your test installation is to your actual network design and hardware (plus some additional resources for the log files generated during the pilot project), the more valuable your results will be as you plan the deployment of SMS throughout your entire corporate network.