Supported Topologies

This topic describes two single-forest Exchange topologies and one multi-forest Exchange topology. These topologies are recommended for minimizing the delay between the time a user makes a change and the time that the change is visible to all other users. This delay is known as free/busy latency. The longer the latency, the more probable it is that users will see out-of-date free/busy data.

These topologies have some free/busy latency because of the client applications. For example, Outlook publishes free/busy data one time every 15 minutes. MadFB checks for free/busy data published by Outlook Web Access or Outlook Mobile Access at 15-minute intervals. Therefore, in any topology, a particular user's free/busy data may be up to 15 minutes old.

Server Per Site Topology

In the server per site topology, each site has a free/busy server that supports the local users. This topology supports one Active Directory forest (and one Exchange organization) that has been subdivided into one or more sites or administrative groups.

Simple topology with one free/busy server per site or administrative group

Users in each site use a local free/busy server. Free/busy data replicates between the different servers. This arrangement works well for topologies that include poorly connected sites (for example, multiple sites that communicate using low-bandwidth connections). Assuming that most users require free/busy data for other users in their own site (for example, for local in-person meetings), this topology provides users with quick access to the most important data.

In cases where a user requires free/busy data for a user in a remote site, latency may become a factor. The free/busy data on one server may be up to 15 minutes out of synchronization with the other servers.

The factors that contribute to the free/busy latency in this topology are the following:

• P   The free/busy publishing interval for Outlook or MadFB. The default value for Outlook is 15 minutes.

• R   The free/busy data replication interval. The default value is 15 minutes.

• F   The inter-forest replication interval. This factor is not relevant for this topology.

For this topology, the maximum delay is P + R, which is one hour if you use the default settings.

Single Server/Multiple Sites Topology

In the single server/multiple sites topology, only one site has a free/busy server. This server supports users in remote sites in addition to the local site. This topology supports one Active Directory forest (and one Exchange organization) that has been subdivided into one or more sites or administrative groups.

Simple topology with one free/busy server for all sites

Users in each site use a central free/busy server. This arrangement eliminates the latency produced by the public folder replication interval, although the remote users may experience some network latency. This arrangement works well for topologies in which all sites have high-bandwidth connections.

The factors that contribute to free/busy latency in this topology are the following:

• P   The free/busy publishing interval for Outlook or MadFB. The default value for Outlook is 15 minutes.

• R   The free/busy data replication interval. This factor is not relevant for this topology.

• F   The inter-forest replication interval. This factor is not relevant for this topology.

For this topology, the maximum delay is P, which is 15 minutes if you use the default settings.

Multiple Forest Topology

In a multiple forest topology, selected Exchange servers replicate free/busy data between Active Directory forests (and the corresponding Exchange organizations). The following figure shows an example of a multiple forest topology where the different forests have different internal topologies. This topology is common in large enterprises.

Complex topology that includes multiple forests that have different free/busy server configurations

A separate Exchange Server 2003 tool, named the Inter-Organization Replication tool (also known as Exchsync.exe), replicates free/busy data from one organization to another. The Inter-Organization Replication tool normally replicates data at 30-minute intervals, which can significantly affect the latency of free/busy data.

The factors that contribute to free/busy latency of this topology are the following:

• P   The free/busy publishing interval for Outlook or MadFB. The default value for Outlook is 15 minutes.

• R   The free/busy data replication interval. The default interval is 15 minutes.

• F   The inter-forest replication interval. The default interval is 30 minutes.

For this topology, the maximum delay is P +nR + mF, where m is the maximum number of inter-forest replication hops required to distribute the replication message, and n is the maximum number of internal replication hops required to distribute the message within an organization before or after the inter-forest replication.

For more information about linking multiple Exchange organizations, see Exchange Server 2003 Deployment Guide.

Inter-Forest Replication Process

Exchange users from different forests can share their free/busy data with each other. There are two primary mechanisms that provide this function:

• Directory synchronization   Ensures that users in one forest appear as contacts in another forest. For a user to retrieve free/busy data for another selected user, that user must appear in the global address list of the local forest.

• Free/busy replication   Pushes free/busy information from one Exchange organization to another. The inter-organizational free/busy replication process uses the MAPI protocol, which means that this is not an Internet-friendly solution.

  Note

  The inter-organizational replication process is one-way. To establish two-way replication between two organizations, you must configure replication separately for each organization.

As described previously, the Inter-Organization Replication tool replicates free/busy data from one organization to another. The tool uses the same MAPI synchronization APIs that Outlook uses to synchronize offline folders. When replicating, the tool uses a unique attribute known as a source key to match data in the source folders with data in the destination folders. The replication process involves the following steps:

1. The tool identifies a list of changes that have to be exported from the source public folder store and identifies the current known state of the destination store.

2. The tool propagates the changes to be processed to the destination public folder store (in the destination organization).

   Note

   If the tool cannot create the destination item in the destination folder, it will delete the source item to maintain synchronization. If the item with that source key has been deleted, and the public folder store has tombstoning enabled, the item cannot be created again and replication will fail.

During normal operation, the Inter-Organization Replication tool generates event log messages such as the following:

Event Type: Information

Event Source: Exchsync

Event Category: None

Event ID: 107

Date: 9/19/2001

Time: 9:12:27 AM

User: N/A

Computer: RED-FBS-05

Description:

Statistics for Free/Busy replication session 'Dogfood --> ITG Free/Busy sync' [81 Message Changes, 0 Message Deletions, 690 Message Ignores, 0 Message Errors, 0 Folder Changes, 0 Folder Deletions, 0 Folder Ignores, 0 Folder Errors].

You can use this information to monitor the status of the replication traffic.