Planning and Designing IIS

Two of the most important tasks to accomplish before implementing IIS are thorough planning and designing. Planning and designing are the beginning phases to properly implementing IIS, and they may consist of the following:

• Defining goals and objectives of the project

• Identifying and reviewing IIS application types and requirements

• Designing the IIS infrastructure to support the goals and objectives

• Designing the back-end infrastructure such as the database or application tier

• Defining security requirements to meet the goals and objectives and balancing the security methodologies between risks and end-user experience

• Examining and designing disaster recovery plans, and monitoring requirements and maintenance practices

• Documenting the current IIS infrastructure and the IIS design decisions

Determining Server Requirements

Hardware and software requirements are based on the information gathered and the requirements set forth in the design and planning stages. The necessary hardware and software requirements should match the goals and objectives of the project. These details are very specific and describe all the resources needed for hardware and software. For example, four IIS servers will each require dual processors, 1GB RAM, triple-channel RAID controllers, and 15K rpm disk drives.

Determining Fault Tolerance Requirements

Fault tolerance is a key aspect of any Web infrastructure and should be addressed during the planning and designing phases. Although some Web sites can afford to have downtime, others may require 99.999% uptime. Service Level Agreements (SLAs) should be determined from the operational goals. When an SLA is in place, such as a minimum of two hours of downtime at any one time, the appropriate fault tolerance can be applied to the Web infrastructure.

Various technologies can be applied to a Windows Server 2003 Web infrastructure to support even the most demanding SLAs. For example, Windows Server 2003 Web servers can use network load balancing (NLB) to distribute the load among multiple Web servers and also provide fault tolerance. NLB is more suited and less costly than using Microsoft Cluster Service to provide fault tolerance. Another way to promote fault tolerance is to tier the environment so that various services are segmented (for example, IIS tier, application tier, database tier, messaging tier), as shown in Figure 11.1.

Figure 11.1. Fault tolerance using a tier environment.