Server locations pull double duty in Exchange 5.5. They alter the algorithm by which clients connect to public folders, and they also let you use the Location address-space restriction. These goals aren't always compatible. You can create server locations that correspond to geographic proximity to address network issues with client access to public folders. For example, if you have a site that spans slow links, the default configuration of replicated public folders can cause clients to randomly select a replica on a server across the slow link. Using server locations, you can control clients so that they have to attempt to connect to all replicas in the server location before they connect to replicas in another server location. In another instance, you might want to create server locations along functional boundaries (e.g., management) that have nothing do with geography. Perhaps in the future you will be able to separate these functions and gain more precise control over how users access connectors in an organization, regardless of public folder access.