The two primary reasons for scaling out a server farm are to achieve high availability and performance. For example, those looking for high availability need at least two Web front-end (WFE) servers, two application servers, and a clustered SQL Server back end. If performance is strictly the goal, you may have only a single four CPU SQL Server server with one very fast WFE server and one application server. Whatever your goal, it is important to note that the new versions of Windows SharePoint Services and Microsoft Office SharePoint Server do not have as many limitations regarding scaling and topology choices as earlier versions. But this increased flexibility in server farm design may make it difficult for some administrators to design an appropriate physical and logical farm topology. If you are not sure how to design your specific server farm architecture, begin with a topology discussed in this chapter and simply change it to meet your needs. This chapter presents four examples of farm topologies: small, medium, large, and enterprise. Due to search limitations, Enterprise farms are covered only as SharePoint Server farms in this chapter.
Remember that these topologies are not concrete rules on which to build your specific implementation; they are simply real-world suggestions about where to begin. Most medium-scale and larger organizations would do well to begin with at least two WFE servers, an application server, and a clustered SQL Server implementation.