Previous Page TOC Next Page



— 14 —
Connecting to Other Exchange Sites


by Kimmo Bergius

After you decide on the placement of sites and servers in your organization, you have to start thinking about connecting your sites. If you are going to have only one site, the connections between the servers are automatic and are configured automatically during the installation of the server software. If you are going to have several sites, you have to choose a connector or connectors to connect these sites together.

Message transfer and communication between two sites differ from message transfer and communication between servers within a site. Servers in the same site communicate with each other using Remote Procedure Calls (RPCs). All communication within a site is server-to-server communication, with all the servers communicating directly with the target server. Two services on each server are responsible for communicating information to servers in the same site: the Directory Service (DS) and the Message Transfer Agent (MTA). The Directory Service is responsible for replicating the contents of the directory between the servers, and the Message Transfer Agent is responsible for transferring messages. All other Exchange services rely on these two services to transfer their information between servers. Communication within a site is also known as intrasite communication.

Communication between sites is called intersite communication. Intersite communication is handled by connectors. There are two different connector types: a messaging connector and a directory replication connector. A messaging connector handles all message traffic between two sites. You can use four different connectors for this task. A directory replication connector is responsible for replicating directory information between sites. The directory replication connector performs this job by placing the information to be replicated into a message and then relying on the messaging connector to transfer the information between the sites. Directory replication and directory replication connectors are discussed in detail in Chapter 15, "Configuring Directory Replication."

A connection between two sites in the same organization consists always of two similar connectors, one in the local site and one in the remote site. In some cases, you can configure both connectors at the same time, but mostly the two connectors must be configured separately in both sites. Both connectors must, however, be configured before a connection exists between the two sites.



If configuring both sites at the same time is possible, you should do this, because this makes the configuration process very simple and error-free.



In this chapter, the term local site is used for your own site, and the term remote site is used for the site to which you are connecting the local site.


Different Messaging Connectors


There are four different connectors that you can use as a messaging connector between two sites. The messaging connectors are


Choosing a Connector


You can choose from four different connectors to use as a messaging connector between two sites. The messaging connectors are as follow:

People have several different opinions about the basic rule for choosing a connector. According to the first opinion, you should always try the site connector first if you have a connection that supports it and then switch to some other connector if the site connector doesn't work properly. The other opinion promotes the use of the X.400 connector over the site connector and suggests that you should always use the X.400 connector if the connection between the two sites is not a direct LAN grade connection. Both of these statements contain a shade of truth. Remember that the selection of a connector between two sites is not final; you can always change the connector if the one you choose does not meet your requirements.



Here's a rule of thumb: When you choose a connector, always start with the site connector if two conditions are met:



If the site connector takes up too much of the connection bandwidth or you want to control the traffic, then switch to the X.400 connector. Use the dynamic RAS connector only if you have no other connection to a remote site than a dial-up connection or as a backup connector, and use the Internet Mail Connector only as a backup connector.

If the preceding tip gave you enough information about choosing a connector, you can go directly to the configuration instructions for your chosen connector later in this chapter. If, however, you want to know more details about choosing a connector, read on.

Before you can choose a suitable connector, you need to find out certain information. Primarily, this information has to do with the type, protocol, and available bandwidth of the connection between the sites, but you should also try to estimate how much network traffic your configuration will create. When you designed your Exchange environment, you probably gathered a lot of information on users and the amount of mail traffic they currently create. You can also use this information to estimate the amount of intersite traffic your users are going to create.

Network Types and Protocols


Before you can choose a connector between two sites, you have to determine the type of the connection between the sites and the protocols that the connection supports.

Network connections can be divided into three categories, depending on the bandwidth the connection offers:


Bandwidth

Connector possibilities

Low-speed connection

Less than 64 kbits/s



Modem or ISDN

Dynamic RAS Connector


X.25

X.400 Connector


TCP/IP network

X.400 Connector,


connection

Internet Mail Connector

Medium speed connection

64 kbits/s to

Site Connector


2 Mbits/s,



Network



connection



supports RPCs



TCP/IP network

X.400 Connector,


connection,

Internet Mail Connector


X.25

X.400 Connector

High-speed connection

More than 2 Mbits/s,


(LAN grade connection)

Network

Site Connector


connection



supports RPCs



X.25

X.400 Connector



Available Bandwidth


You can always determine the total bandwidth of the network connection between two sites. Determining how much of this bandwidth is actually available is much more difficult. Available bandwidth means how much of the total bandwidth is left after the bandwidth used by other applications using the same connection has been deducted.

You might be using a 64 kbit/s leased line, for example, to connect two Exchange sites. Users could also be accessing a database or transferring files over the same connection. This means that the connector between the two sites would no longer have the whole 64 kbit/s to itself but must coexist with the other network traffic. To calculate available bandwidth, you must be aware of all other traffic using the same connection.



You can monitor the usage of your network using any network monitoring software, such as the Network Monitor included in Microsoft System Management Server. You can also use NT's Performance Monitor to monitor the amount of available bandwidth.


Bandwidth Used By Exchange


To determine whether your existing network will support Exchange, you should try to estimate how much network traffic will be created by your Exchange environment. The following factors affect the network traffic in the Exchange environment:

You can estimate many of these factors by analyzing the traffic created by users in your existing system. Ask different users to monitor their mail system usage for a few days, and record the number of messages sent to the same system, to foreign systems, and to public folder usage. This information will tell you a lot about the initial requirements of your new system and will help you plan your hardware, software, and infrastructure requirements. If you are currently using a mail system based on file sharing, and the system performance is adequate, the same network infrastructure will probably support migrating to an Exchange system. Keep in mind, though, that as the new system offers better performance and new capabilities, the usage of the system will increase, thus increasing your hardware and infrastructure requirements.

Multiple Connectors


You can define multiple connectors between two sites. The connectors can be all of the same type or of different types. You use multiple connectors for two reasons:

Fault tolerance means that you can have two or more connectors between two sites usually using different connections. Having two or more connectors ensures that if one of the connections and connectors is unavailable for some reason, the message can be sent using another connector.

Load balancing means dividing site-to-site message traffic between two different connectors. In load-balancing configurations, the connectors are usually of the same type, and they can use the same or different connections. If you have a lot of message traffic between two sites, and the connection is not the bottleneck, you can divide the processing of messages between two servers in both sites. You therefore would configure two connectors between the sites into two different servers. The cost of the connectors participating in load balancing must be the same.

Connector Cost


If multiple connectors are used between two sites, the MTA on the local server must be able to determine which one should be used to transfer the message. The MTA uses a definition called the cost to determine which connector should be used. The cost is defined as a number from 1 to 100. If multiple connectors are used between two sites, the one with the lowest cost associated with it is used to transfer the message.

Connector costs are cumulative, which means that if a message has to travel through multiple connectors, the cost from the source server to the target server is the cumulative cost of all the connectors handling the message. If you have three sites—for example, A, B, and C—A and B are connected using a site connector with a cost of 1. B and C are connected with a similar site connector with a cost of 1, and A and C are connected with an X.400 connector with a cost of 5 (see Figure 14.1.). Messages from A to C are routed through site B because of the lower cost (1 + 1 = 2 compared to 5).

Figure 14.1. Multiple routes.

What happens if a message is sent from site A to site C, and the connector from site B to site C is unavailable? The MTA on the server on site A still sends the message to site B because it cannot be aware of the status of the connector between sites B and C. Site B also calculates a route for the message, but it knows that the connector between B and C is not available, so it uses the least costly route to C, which is via A. The MTA on A again checks the routing and finds out that the cheapest way from A to C is via B, but it also sees that the route via B has already been used, so it sends the message along the cheapest available route, which is directly to C.

How Does Exchange Select a Connector?


When you send a message that is addressed outside your home server, the MTA on your home server examines the recipient address of the message. If the message is to another server in your own site, the MTA establishes a connection directly to the MTA on the target server and transfers the message.

If the recipient of the message is not within your own site, the MTA examines the possible routes to the target server and chooses the route that has the lowest cost associated with it. The cost of a route from the local site to the remote site is the cumulative cost of all the connectors through which the message has to pass. After the MTA chooses the lowest cost route, it passes the message to the first connector in this route, and this connector is responsible for transferring the message to the next site in the route.

Property Pages Appearing in All Connectors


Some property pages are common to several connectors. These property pages are explained in the following sections. Each property page also has a list of connectors for which you can find the property page.

Address Space



Applies to:

Site Connector


X.400 Connector


Dynamic RAS Connector


Internet Mail Connector


(MS Mail Connector)


All connectors, whether they are used to connect two Exchange sites or an Exchange system and another mail system, have to be associated with an address space. The address space defines a set of addresses that the connector is to handle. You could define an X.400 connector, for example, to connect an Exchange site to another site in the same organization and specify the following value for the connector's address space:

X400

c=us, a= , p=organization_name, o=site_name


where organization_name and site_name are the organization and site name of the remote site, respectively. This address space defines that the X.400 connector will handle all sent messages in which the X.400 address fields c, a, p, and o match the values defined in the address space.

You define the address space by selecting the Address Space tab in the Connectors property window. The program displays the Address Space property page shown in Figure 14.2.

Figure 14.2. The Address Space property page.

An address space definition should contain the fields that are necessary to distinguish the messages that the connector should handle. You can also define multiple address spaces and even multiple address types for a single connector.



You can use an asterisk (*) as a multicharacter wildcard and a question mark (?) as a single character wildcard in an address space.

In an X.400 address, the Administrative domain field (ADMD, a field) cannot be left blank. If you want to define that no ADMD is needed, type a space in the field.



Connector Schedule



Applies to:

X.400 Connector


Dynamic RAS Connector


You can set the X.400 connector or dynamic RAS connector to initiate a connection to the remote server at specified times by defining a schedule. You define the connector schedule by selecting the Schedule tab in the Connectors property window. The program displays the Schedule property page shown in Figure 14.3.

Figure 14.3. The Schedule property page.

You can choose from four options to determine the schedule:

Remote Initiated

The local connector never initiates a connection to the remote site but rather waits until the corresponding connector in the remote site initiates a connection. After the remote connector finishes sending messages to the local site, the local connector sends all queued messages to the remote server.




You can define the Remote Initiated option in only one of the connectors participating in a connection. If the schedule in both connectors is defined as Remote Initiated, no messages are ever delivered.

Never

This option disables the connector.

Always

The connector establishes a connection to the remote site approximately every 15 minutes and sends all queued messages.

Selected times

The connector operates according to the schedule defined in the bottom of the window. You can specify the schedule using either a 1-hour or a 15-minute schedule grid.


You cannot define a schedule for the site connector or the Internet Mail Connector. These connectors assume that the network connection between the sites is permanent and always available.

Delivery Restrictions



Applies to:

X.400 Connector


Dynamic RAS Connector


Internet Mail Connector


Delivery restriction means that you can allow or deny the right to use the connector on a user basis. You set delivery restrictions by selecting the Delivery Restrictions tab in the connector's property window. The program displays the Delivery Restrictions property page shown in Figure 14.4.

Figure 14.4. The Delivery Restrictions property page.

You can allow only specified users to use the connector by adding the names of these users in the Accept Messages From list. Alternatively, you can deny the use of the connector from specified users by adding the names of these users in the Reject Messages From list. In this case, everyone else except the users listed will have access to the connector. If a user is not allowed to use a certain connector, the system informs of this restriction by sending an error message to the user.

Connected Sites Property Page



Applies to:

X.400 Connector


Dynamic RAS Connector


Internet Mail Connector


If the messaging connector used to connect the two sites is an X.400 connector, dynamic RAS connector, or the Internet Mail Connector, you have to define the remote site in the connector's Connected Sites property page, as shown in Figure 14.5. Otherwise, the directory replication connector cannot use the messaging connector to transport the replication messages. Only the site directly in the other end of the messaging connector has to be defined on the Connected Sites page. All the other sites within the organization are defined automatically through the Knowledge Consistency Check. You do not have to define the connected sites for the site connector because this information is configured automatically.

Figure 14.5. The Connected Sites property page.

Restricting Message Size


You can restrict the size of messages delivered through an X.400, dynamic RAS, or Internet Mail Connector by specifying the maximum message size in kilobytes. The specified size is the size of the total message. You cannot define a maximum size for attachments. You define the maximum message size on the Advanced property page for the X.400 connector and on the General property page for the dynamic RAS and Internet Mail Connectors. You cannot define the maximum message size for the site connector.

Site Connector


The following table lists the advantages and disadvantages to using the site connector:

Advantages

Disadvantages

Most efficient of the connectors.

Requires a fast and permanent network connection that supports RPCs.

Easiest connector to configure; both the local site and the remote site can be configured at the same time.

You have relatively little control over the operation of the connector, such as connection scheduling, maximum message size, and users allowed to use the connector.

Automatic load balancing and fault tolerance through multiple source and target servers.

The connector can saturate a link if multiple servers in the local site establish a connection to multiple servers in the remote site at the same time.


You should choose the site connector if

The site connector is the most efficient of the connectors because it does not convert messages from one format to another. Instead, the message stays in the internal Exchange format. For this reason, this connector is sometimes called the Exchange connector.

The site connector uses RPCs to communicate information between servers in two different sites. This is another reason for the efficiency of the site connector. This factor is also limiting, however, because the use of RPC requires a protocol that can support RPC.

When you install and define the site connector, no new connector is installed in the system. The site connector is just a set of definitions that tell the Message Transfer Agent that it should transfer messages to the other site as well as to servers within the same site.

Target Server and Bridgehead Server


A target server is the server in the remote site to which the server in the local site establishes a connection to deliver the message. All servers in the remote site can be target servers. A bridgehead server is a server in the local site that takes care of all the message traffic to the remote site. By default, no bridgehead server is defined for the local site, but instead all servers within the local site can establish a connection to all servers in the remote site. By default, all the servers within the remote site are defined as target servers.

As the administrator, you can direct message traffic to the remote site through a specific server in the local site by selecting one of the servers to act as a bridgehead server. Only the bridgehead server is therefore allowed to establish a connection to the remote site. All other servers in the local site transfer a message destined to the remote site first to the bridgehead server, which then initiates a connection to a server in the remote site and transfers the message.

You can also limit the number of servers in the remote site to which the bridgehead server (or all servers, if no bridgehead server has been defined) in the local site can establish connections. You do so by using the target server's property page. Initially, all the servers in the remote site are listed as target servers, but you can remove servers from the list and thus direct message traffic through certain target servers.

If you are defining a site connector between two sites and want to direct all traffic between one server in the local site and one server in the remote site, you should configure the connector in both sites so that a single server in the local site is selected as a bridgehead server and a single server in the remote site is defined as a target server.

All target servers are associated with a cost, which defines the selection of the target server. The cost is defined as a number from 0 to 100. The costs 0 and 100 have special meanings: 0 means that the specified target server is to be used always if it is available, and 100 means that the target server is used only if no other target servers are available. The costs 1 to 99 are used to determine the target server using a weighed average calculation. This means that if one target server has the cost 1, and another has the cost 2, the first one is used roughly twice as much as the second one.



Do not confuse the target server cost with the cost specified for a connector. The target server cost is defined as a number from 0 to 100, and the connector cost from 1 to 100. Also, the target server cost will only affect how much each target server will be used (excluding the special cost values 0 and 100), whereas the connector cost will actually determine the order in which the connectors will be used.

Figure 14.6 illustrates three different site connector configurations between two sites. Both the sites have three servers: A, B, and C. In the first example, the site connector is defined using default settings; that is, no bridgehead server in the local site and all servers in the remote site as target servers. In the second example, a bridgehead server has been selected in the local site, but all the servers in the remote site are still target servers. The third example shows how message traffic between the two sites has been directed through two servers only, where one server in both sites has been selected the bridgehead server and just one server in the remote site has been defined a target server.

Figure 14.6. Different site connector configurations.

Site Service Account


Site connector uses RPCs to transfer the information between sites much in the same way as two servers within the same site communicate with each other. Also, the service account used to start the server MTA services affects the operation of a site connector. If the site service account used in the local site has no access rights to the remote site, the site connector does not work properly.

The easiest configuration is to have the same service account in both sites. This configuration is easy to establish if the sites are within the same NT domain. Because the site connector requires a permanent connection between the sites, it is also possible and quite feasible to set up a trust between the sites and use the same service account in both sites, if the sites are within different NT domains.

Sometimes you cannot use the same service account in both sites. For these cases, you can override the local site service account and define a different account to be used when connecting to the target site. You do so by using the Override property page, as shown in Figure 14.7. The Override feature is much the same as the File Manager Connect As feature you use when connecting to network resources in different domains when no trust exists between the domains.

Figure 14.7. The Override property page.

Installing and Configuring the Site Connector


The following steps include only the necessary definitions to install and configure the site connector. Additional definitions are discussed briefly after the basic installation steps.

  1. Open the Exchange Administrator program, and choose File | New Other Site Connector.



  2. The program might display a message informing you that the connector cannot be installed in the current container and asking whether you want to switch to the Connections container. Click OK in the message box.

  3. Type the name of a server in the remote site. The server can be any of the servers in the target site. The Site Connector Properties dialog box appears displaying information for the remote site.

  4. Click OK. The program displays a message asking whether you want to create the site connector in the remote site as well.



  5. You can create the site connector for both sites at the same time only if you have access rights to the remote site and the selected server. Otherwise, you must create the site connector in each site separately.

  6. Choose Yes if you want to create the site connector in the remote site as well. Otherwise, choose No. The Site Connector Properties dialog box appears.

  7. Click OK.

If you chose to create the site connector to the target site in step 4, the sites are now connected, and you can test the connector by sending a message to a user in the remote site. Otherwise, you must configure the remote site separately before the site connector configuration is finished and you can start sending messages to the remote site and back or configure a directory replication connector between the sites.



Always test the site connector before going further by sending a message to a user in the remote site using the user's full X.400 address.


Advanced Configuration of the Site Connector


The Site Connector property pages contain some additional settings that you can optionally use to further control the operation of the connector.

General Property Page

Apart from the name of the connector and the remote site, you can use the General property page to define a cost for the connector or a bridgehead server in the local site.

The connector cost is used to select a connector to transfer the message between two sites in a multiple connector configuration. Type the cost in the Cost field. For more information, see "Connector Cost" earlier in this chapter.

A bridgehead server enables you to direct message traffic to the remote site through a specific server in the local site. You can define a bridgehead server by selecting Specific Server and then selecting the server name from the list. For more information, see "Target Server and Bridgehead Server," earlier in this chapter.

Target Servers Property Page

On the Target Servers property page, you can define the servers in the remote site to which the servers in the local site can establish a connection. For more information, see "Target Server and Bridgehead Server," earlier in this chapter.

Address Space Property Page

The X.400 address of the target site is automatically entered in the Address Space property page. You can modify this address space or define additional address spaces using the options in this property page.

Override Property Page

A site connector uses RPCs to communicate between the two sites. This capability requires that both sites use the same site service account. If the sites do not have the same service account, however, you can define the service account of the remote site in the Override property page. This task is much the same as using the Connect As option when connecting to a resource in File Manager.

The X.400 Connector


The following table lists the advantages and disadvantages to using the X.400 connector:

Advantages

Disadvantages

Gives you a lot of control over how and when the connector is used and which users can use it.

Not as efficient as the site connector.

Messages to other Exchange sites and organizations can be transferred in the Exchange internal format, which internal format, which improves efficiency.

Can create bottlenecks when a lot of traffic between two sites is routed through two machines.

Supports TCP/IP, TP0/X.25, and TP4/CLNP connections.

Relatively complicated to configure.


You should choose the X.400 connector if

Compared to the site connector, the X.400 connector gives you a lot more control over how and when the communication between two sites is to occur. The X.400 always uses one server in the local site and one server in the remote site to transfer the information. This setup enables you to direct message traffic through certain servers. You can also define a schedule for the connector, as well as limit the maximum size of messages transferred over the connector, and allow or deny access to the connector on a user basis.

When you configure an X.400 connector between two sites, one server in both sites is chosen to act as the messaging bridgehead server for this connector, and the connectors are configured on these servers. This setup directs all message traffic between the two sites through these servers.

The X.400 connector enables you to connect two sites using an X.400 network as a backbone. You therefore do not need to have your own connection between two sites, but instead both sites need to have a connection only to a public X.400 network, which can be used as a backbone. You can also connect two Exchange sites directly using an X.25 or a TCP/IP capable network connection.

The X.400 connector supports three different protocols: TCP/IP, TP0/X.25, and TP4/CLNP. Before you can define the X.400 connector, you have to define a suitable transport stack. For most connections, especially connections between two Exchange sites, using the TCP/IP transport stack is easiest. You use the TP0 stack only for connections over an X.25 network and the TP4 stack only if you are configuring a connection to a mainframe computer that supports only the TP4 protocol.

Installing and Configuring the X.400 Connector


The following steps cover the installation of the TCP/IP transport stack, which is the most common of the transport stacks. Before you can define the transport stack, you have to install TCP/IP and configure it on the server.

  1. Open the Exchange Administrator program, and choose File | New Other | MTA Transport Stack. The program displays a window with a list of available transport stacks.

  2. Select the TCP/IP MTA Transport Stack and the server where the stack is to be installed. Then click OK.

  3. The program displays the property window for the TCP/IP transport stack. On the General page, you can define a name for the transport stack (or accept the default name). You need to define the OSI address information only if multiple applications or services are using the same transport. The Connector page shows all connectors using this transport stack.

  4. Click OK.

The transport stack has now been configured, and you can go on by configuring the X.400 connector. The following steps include only the necessary definitions to install and configure the X.400 connector. Additional definitions are discussed briefly after the basic installation steps.

  1. Choose File | New Other | X.400 Connector. The program displays a window displaying all the possible X.400 connector types. Select TCP/IP X.400 Connector, and click OK. The program displays the Properties window for the X.400 connector.



  2. The program might display a message informing you that the connector cannot be installed in the current container and asking whether you want to switch to the Connections container. Click OK in the message box.

  3. Type a Display and Directory name for the X.400 connector in the corresponding fields. Use names that enable you to identify the connector in the Connections container—for example, X400 CONNECTOR TO SITE3.



  4. Some of the definitions within the Exchange server are case-sensitive. You therefore should always choose to type all names and other definitions in either upper- or lowercase. Generally, using uppercase letters is advisable.

  5. Type the Remote MTA name and password in the corresponding fields. When you are configuring an X.400 connector to another Exchange site, the Remote MTA name is the name of the remote server, if you have not defined another name on the Override page of the remote server's X.400 connector (for more information, see "Override Property Page" later in this chapter). Also, by default, the password has not been defined unless you have defined a password for the remote server's MTA, either in the remote server's Message Transfer Agent property window or in the X.400 connector's Override page.

  6. Select the Schedule tab, and define a schedule for the connector. For more information, see "Connector Schedule" earlier in this chapter.

  7. Select the stack page. Select IP Address, and type the IP address of the remote in the field.

  8. Select the Address Space tab, and define an address space for the connector by clicking New X.400. The program displays a window, where you should type the following information:

  9. Field

    Type

    Organization (o)

    Site name of the remote site

    Private management domain name (p)

    Name of the Exchange organization

    Administrative management domain name

    A single space

    Country (c)

    Select your country from the list


  10. Type a cost for the connector in the Cost field (for more information, see "Connector Cost" earlier in this chapter). Then click OK. The defined address space appears in the list.

  11. Select the Connected Sites tab. Select New, and type the name of the remote site in the Site field. Then click OK. The remote site then appears in the Connected Sites list. For more information on connected sites, see "Connected Sites" earlier in this chapter.

  12. Click OK in the property window. The X.400 connector is now defined for the local site. For the connection to work between the sites, you have to configure the X.400 connector in the remote site by performing the same steps.



Always test the X.400 connector before going further by sending a message to a user in the remote site using the user's full X.400 address.


X.400 Connector Advanced Configuration


You can further fine-tune the X.400 connector by using the properties explained in the following sections. Some of the property pages (Schedule, Delivery Restrictions, Connected Sites, and Address Space property pages) are also discussed earlier in this chapter.

Override Property Page

If you want to override the local MTA name and password defined in the local server's Message Transfer Agent property page, you can define connector-specific values on the Override property page. Connector-specific passwords are handy in some cases, because if you have to remove the connector for some reason, you do not have to change the passwords on all the other connectors for security reasons. The MTA name and password entered for the X.400 Connector in the remote site have to exactly match the values entered here.

You can also use the Override property page to define connector-specific connection values. Values specified here override the values specified in the Messaging Defaults property page in the MTA Site Configuration Properties window. You do not generally need to alter these values.

Advanced Property Page

On the Advanced property page, you can specify X.400 specific settings, such as the X.400 mode and link options and maximum message size. The X.400-specific options are defined by the MTA to which the X.400 connector is connected. If you are connecting two Exchange sites using an X.400 connector and a direct TCP/IP connection between the sites, you should make sure that the settings 1988 normal mode, Allow BP-15, and Allow MS Exchange contents are selected. These settings define that the X.400 connection conforms to the 1988 X.400 standard and that messages can be transferred using the Exchange internal message format (the most efficient X.400 mode).

The Two-Way Alternate setting enables you to define whether incoming messages from the remote site will be transferred during a connection established from the local site. This setting is particularly useful when the connection between the two sites is an expensive X.25 network. Thus, each site is responsible only for costs associated to message traffic initiated in the local site.

You also can use the Advanced property page to define the maximum size of messages transferred over the X.400 Connector. If you want to limit the maximum messages size, select Maximum (K) and type the required size (in kilobytes) to the field.

Dynamic RAS Connector


The following table lists the advantages and disadvantages to using the dynamic RAS connector:

Advantages

Disadvantages

Gives you a lot of control over how and when the connector is used and which users can use it.

Data transfer rate is dependent on the speed of the modem or ISDN line.

Supports asynchronous modem and ISDN connections.

Can create bottlenecks when a lot of traffic between two sites is routed through two machines. Relatively complicated to configure.


You should choose the dynamic RAS connector if

You can use the dynamic RAS connector to connect two sites much in the same way you use a site connector. The biggest difference is that this connector does not require a permanent connection to the remote site. Instead, the dynamic RAS connector can use connections defined in the NT Remote Access Service (RAS) to communicate to the target site.

With RAS, you can connect two NT server computers using a normal telephone line (modem connection), an ISDN line, or an X.25 connection. The dynamic RAS connector can use any defined RAS connection. The RAS connector is used to connect two sites if no permanent connection exists between the sites.

Before you can install and configure the dynamic RAS connector, you have to install and configure the Remote Access Service on your NT server. You do so using the Network icon in the Control Panel. You should also define a phone book entry for the remote server to which you want to connect. The RAS phone book entries are configured using the Remote Access program (which you can find in the Remote Access Service program group).

The following steps illustrate the configuration of the RAS MTA Transport Stack:

  1. Choose File | New Other | MTA Transport Stack. The New MTA Transport Stack window appears.

  2. Select RAS MTA Transport Stack from the Type list box and the server on which the stack is to be installed from the Server box. Click OK. The RAS Properties dialog box then appears.

  3. The name of the transport stack is automatically entered in the Name field in the General property page. Optionally, you can define an MTA callback number for the local transport, which can be used to further enhance data security when using a RAS Connection. If callback is enabled, the server initiates a connection to the remote server, and after the remote server answers the connection, is terminated. The local server waits until the remote server calls back to the number defined here. The callback requires that RAS callback security is enabled in RAS.

    In the Connectors property page, you can see a list of connectors that use this transport stack.

  4. The RAS transport stack is now configured, and you can go on configuring the dynamic RAS connector. The following steps include only the necessary definitions to install and configure the dynamic RAS connector. Additional definitions are discussed briefly after the basic installation steps.

  5. Choose File | New Other | Dynamic RAS Connector. The program displays the RAS Connector Properties window.



  6. The connector is installed in the Connections container. If you are not in this container, the program displays a dialog box asking whether you want to move to this container. Click OK.

  7. In the General property page, type in the display and directory names for the connector. Also type the name of the remote server, and select the transport stack to be used and the RAS phone book entry. You can define a new phone book entry by selecting RAS Phone Book.

  8. Select the Schedule tab. In the Schedule property page, you can define the times when local server can initiate a connection to the remote server. For details, see "Connector Schedule" earlier in this chapter. By default, the schedule is set to Always.

  9. Select the RAS Override tab. On the RAS Override property page, you should define a user account that has access rights on the target server. The account should be defined at least as a user (have at least Send As and Mailbox Owner rights) on the Servers and Configuration containers in the target server. You can also specify the site service account for the remote site. On this page, you can also override the call and callback telephone numbers.

  10. Select the Address Space tab. Define an address space for the connector by clicking New X.400. The program displays a window, where you should type the following information:


  11. Field

    Type

    Organization (o)

    Site name of the remote site

    Private management domain name (p)

    Name of the Exchange organization

    Administrative management domain name (a)

    A single space

    Country (c)

    Select your country from the list


  12. Type a cost for the connector in the Cost field (for more information, see "Connector Cost" earlier in this chapter). Then click OK. The defined address space appears in the list.

  13. Select the Connected Sites tab. Select New, and type the name of the remote site in the Site field. Then click OK. The remote site then appears in the Connected Sites list. For more information on connected sites, see "Connected Sites" earlier in this chapter.

  14. Click OK in the property window. The Dynamic RAS Connector is now defined for the local site. For the connection to work between the sites, you have to configure the connector in the remote site by performing the same steps.



Always test the dynamic RAS connector before going further by sending a message to a user in the remote site using the user's full X.400 address.


Dynamic RAS Connector Advanced Configuration


You can further fine-tune the Dynamic RAS Connector by using the properties explained in the following section. Some of the property pages (Schedule, Delivery Restrictions, Connected Sites, and Address Space property pages) are also discussed earlier in this chapter.

MTA Override Property Page

You can use the MTA Override property page to define connector-specific connection values. Values specified here override the values specified in the Messaging Defaults property page in the MTA Site Configuration Properties window. You do not generally need to alter these values.

Internet Mail Connector


The following table lists the advantages and disadvantages to using the Internet Mail Connector:

Advantages

Disadvantages

Gives you a lot of control over the maximum message size and which users can use it.

Messages have to be converted from the Exchange internal format to the SMTP format, which decreases performance.

Supports existing TCP/IP and SMTP backbone networks.

Can create bottlenecks when a lot of traffic between two sites is routed through two machines. Connections cannot be scheduled.


You should choose the Internet Mail Connector if

Using the Internet Mail Connector, you can connect two Exchange sites within an organization and then use the Directory Replication Connector over the IMC. You can connect the sites, for example, over the Internet, but doing so is not recommended for the following reasons:

Control: Because the Internet is a collection of networks administered by many different companies and institutions, none of which is responsible for the whole Internet, you have no control over whether the link is available.



If you have an existing permanent TCP/IP connection between two sites, in most cases you should use the X.400 connector to connect the sites. The X.400 connector supports direct TCP/IP connections as well and is much more efficient than the IMC. You should use the IMC as a connector between two sites only in cases where you want to use an existing SMTP backbone network to connect the sites.

This chapter does not provide any information on how you configure the Internet Mail Connector. The procedure is essentially the same as when you use the IMC to connect your Exchange system to another SMTP system, which is discussed in detail in Chapter 13, "Connecting to Other Mail Systems." The only difference when using the IMC to connect two sites within an organization is that you have to define the remote site in the Connected Sites property page so that you can configure the Directory Replication Connector over the IMC.

Summary


Selecting a connector to connect two sites within an Exchange organization is one of the most important and difficult tasks that faces an administrator configuring an Exchange system. The steps needed to perform the selection and configuration are:

The most important thing to remember is that you can always change the connector if it does not suit your needs. Therefore, you should always start with the Site Connector if your connection supports its use. Later you can move on and use the X.400 Connector, if necessary. The Dynamic RAS Connector and the Internet Connector should only be used as backup connectors, or in the case of the RAS Connector, if you only have a dial-up connection between the two sites.

After you have successfully installed and configured the messaging connector between the sites, you can configure the directory replication connector, which is detailed in the following chapter.

Previous Page TOC Next Page