Microsoft Exchange Server 2007: Tony Redmond's Guide to Successful Implementation

464 6.4 Routing ABC

6.4 Routing ABC

Before we look at different routing scenarios, we have enough detail to

present a picture of how the major components link together to establish

message flow in an example Exchange organization. Before we begin, let’s

recap some points:

 The Active Directory Topology Service provides information to the

Exchange transport service about sites and site links. Exchange uses

this data to calculate the least-cost path for a message and to under-

stand whether any hub routing sites occur along the path and if it

needs to bifurcate a message at any point.

 Hub transport servers make point-to-point connections with other

hub transport servers to transfer messages as close as possible to the

final destination.

 If a hub transport server is unavailable in the destination site, then

the message goes to a hub transport server in the closest site (as dic-

tated by routing costs). If Exchange can’t reach a hub transport

server in the closest site, then it “backs off” and attempts to transfer

the message to a hub transport server in the next closest site, and so

on until the message is eventually on its way. If it is impossible to

transfer a message to another site, Exchange queues it until a con-

nection is possible.

 Messages that arrive into an intermediate hub transport server stay

there until Exchange can transfer them to their final destination or to

a hub transport server in a closer site.

Figure 6.14 shows Exchange 2007 servers installed into four Active

Directory sites (Dublin, Paris, London, and Prague). Recall that you don’t

need to configure the send and receive connectors that allow messages to flow

between the hub transport servers between sites as this is done automatically

when you install the hub transport role on a server. Each of the sites support

some mailbox servers and the Dublin site hosts two other connectors: one is

a custom SMTP connector to route messages onwards to a smart host for

delivery to Internet addresses; the other links Exchange 2007 to the legacy

servers in the rest of the organization through a routing group connector. In

this example, Dublin and Paris are hub routing sites. Now that we have a

mental picture of how to link everything together, let’s see how routing works

in practice.

The simplest routing situation for the transport service is where a mes-

sage sent on a mailbox server in Paris is addressed to a mailbox in Dublin.

6.4 Routing ABC 465

Chapter 6

The message travels from the mailbox server to the Paris hub transport server

and is then transferred using a point-to-point SMTP connection over to the

Dublin hub transport, which then delivers it to the mailbox server that hosts

the destination mailbox.

Slightly more complex, a message from a mailbox in London to a mail-

box in Dublin or Prague must travel through the hub transport server in Paris

because it is a hub routing site en route to Dublin. The same is true for mes-

sages generated in Prague, which flow through Paris en route to Dublin or

London. If the Paris hub transport server is unavailable, then messages queue

up because they cannot be transferred. If Paris was not defined as a hub rout-

ing site, then the hub transport server in London could make direct connec-

tions to the hub transport server in Prague or Dublin.

All messages to Exchange 2003 must flow through the Dublin hub

transport server because it is the only hub transport server that is associated

with the routing group connector. All messages flowing in from Exchange

2003 travel through the Dublin hub transport server, which then makes

the necessary connections to Paris to transfer messages that are not

addressed to recipients in Dublin. Remember that Paris is a hub routing

site, so messages from Exchange 2003 that are destined for mailboxes in

Prague and London flow through Paris. The same situation exists for Inter-

Figure 6.14

How the routing

pieces fit together