Chapter 11. Exchange Server 2007 Routing

Routing determines how a message gets from a source server to its destination server. Routing decides on the best and least-expensive path a message takes when transferred between Exchange servers within an organization and to servers in other organizations. In Exchange Server 2000/2003, several components were involved in routing, including Link State, Routing Groups, Connectors, and Routing Group Masters. Routing Groups were logical groupings of Exchange servers determined by the administrator. In Exchange Server 2007, because the transport core functionality has changed as noted in Chapter 7, message routing has also been changed. Rather than using the Routing Engine found in Exchange 2000/2003, the Categorizer (see "The Transport Server Architecture" in Chapter 7) now has a routing stage that determines the ultimate destination for a message, selects a route to that destination, and selects and resolves the next hop for that destination to a server(s) and IP addresses.

Link state was one of the components of routing in Exchange 2000/2003 many Exchange administrators were happy to see go away. Although link state had some benefits, managing it in some cases required a significant overhead. In some large Exchange Server environments, the orgInfo packet, which holds the routing information for the organization, became quite large and caused significant network bandwidth utilization during data transfer among servers. Also, transient network and Active Directory replication problems caused connector oscillations. Rather than use a logical groups of servers, Exchange Server 2007 takes advantage of Active Directory site topology in routing messages within and outside an Exchange organization.