- Mapping the Design Components
- Evaluating Different Design Options
- Active Directory Design Details
- Defining Storage Groups and Multiple Databases
- Defining Administrative and Routing Groups
- Designing Remote Access to Exchange 2000
- Exchange 2000 Support and Maintenance Tasks
- Case Study for SmallCompany Inc.
- Case Study for MediumCompany Inc.
- Case Study for LargeCompany Inc.
- Summary
Defining Administrative and Routing Groups
These two core elements are the building blocks for the Exchange 2000 implementation. Based on the needs of the organization, the criteria used to determine administrative and routing groups will change. However, there are basic principles regarding these two groups that are constant. The administrative model that is presently in place will provide the roadmap for defining the Exchange administrative groups. The Active Directory site replication topology will map directly to the routing group design. However, there are always exceptions. Depending on the specifics of the organization, these two principles may or may not apply.
Determining Administration Delegation
Earlier versions of Exchange used the concept of Sites to establish the boundaries between the physical and administrative topologies of the organization. Servers were grouped and assigned to Sites to control administration. Sites were established based on the constraints of the physical networks and the available bandwidth.
Exchange 2000 has separated this into two distinct categories: administrative groups and routing groups. The following sections describe the different types of administrative models and how they can be applied.
The administrative groups act as the shell for routing groups, servers, public folder trees, policies, and other related objects. However, it functions as more than just a container for the above-mentioned objects. The administrative group allows for the delegation and decentralization of administration for the Exchange 2000 installation. This is achieved through the customization of permissions based on the organization's needs. For example, if the organization is separated into three administrative teams and each team is responsible for managing its own Active Directory domain, it would be reasonable to assume that three separate Exchange administrative groups would be created to manage the servers in each domain.
Servers and routing groups are contained within administrative groups. When planning for server placement within administrative groups, be aware that while in Exchange server mixed mode, once they are defined servers cannot be removed from the specified administrative group. Only when in native mode can servers be moved between administrative groups.
NOTE
Exchange 2000 creates a default administrative group if one is not specified during the installation. In order to specify the administrative group that is used, there must be a group created prior to the install. When building the first installation for the Exchange organization, install the Exchange System Manager tool to create the administrative groups that will be used by the organization. This allows for the specification of the administrative group that the installation uses during the setup procedures.
Distributed Exchange Management
Distributed Exchange management occurs when an organization delegates the infrastructure's administrative responsibilities based on sites or geographical regions. In this configuration, there would be an administrative group assigned by site or geographic region.
Centralized Routing Management
A centralized routing management structure is one in which the management of the Exchange installation falls into the responsibility of a single group of administrators or into the responsibility of a single physical site designation. The concept is similar to that of an organization that would have a centralized data center managing the entire infrastructure. There may be at most one or two defined administrative groups with many routing groups.
Centralized Policy Management
There are different levels and types of policies that dictate how the Exchange servers are managed and monitored. In this model, there are two types of groups. The core group is responsible for creating the policies as they relate to the Exchange installation. In addition, this group can be subdivided and given the responsibility of creating the policies that control the routing group configurations.
The regional IT administrators manage the day-to-day tasks and monitor the overall health of their installations. The regional administrators have limited authority based on the policies defined by the core group. Depending on the level of complexity of the routing topology, the regional IT administrator may or may not be allowed to do more than monitor the message routing activities. If the level is considered to be complex, the core group may allow only monitoring and choose to maintain control.
Routing Groups
A routing group is a method of organizing servers based on the bandwidth constraints of the connections between the servers. The concept behind routing groups is the ability to control mail flow. It also allows for tracking and troubleshooting of messages transferred between groups. Routing groups can be compared to the concept behind the Exchange 5.5 Sites. However, there are important differences. They are as follows:
Servers within a routing group communicate via SMTP. In Exchange 2000, all intra-_routing group communications is done over SMTP, which is different from an Exchange 5.5 installation where servers within an Exchange v5.5 Site communicate over Remote Procedure Calls (RPCs).
Communications between routing groups can now be secured by using the SMTP connector that leverages the Windows 2000 Certificate Server and Exchange 2000 Key Management Services.
Routing, Transport, and Connectivity
Message routing between servers within the same routing group are routed directly from the source server to the destination server. Routing occurs in this fashion because it is assumed that the connections between all the servers within the routing group are high bandwidth and well connected. Within every given routing group there is a defined routing group master. By default, the first server in the routing group becomes the routing group master. If required, this can be manually changed using the Exchange System Manager. All messages with a destination outside of the routing group are referred to the routing group master and then transported via the most cost-efficient route.
Benefits of Going to Native Mode Versus Mixed Mode
Exchange 2000 also has a mixed mode and a native mode. However, this is not to be associated with the Windows 2000 mixed and native mode feature. Under Exchange 2000 mixed mode, Exchange 5.5 servers can coexist and communicate with Exchange 2000 using Remote Procedures Calls (RPC) and Site Connector emulators. However, while in mixed mode, you are restricted from moving servers between routing groups. This can only be accomplished when in native mode. The advantages when operating under native mode are as folows:
The ability to move routing groups, servers, and mailboxes between administrative groups.
The SMTP protocol is set as the default routing protocol.
The same restriction applies when switching from Exchange 2000 mixed mode to native mode; once you switch to native mode, you cannot resort back to mixed mode.
Connectors
There are three Routing Group Connector options available. The options are
Routing Group Connector. The Routing Group Connector is the replacement to the Exchange 5.5 Site connector. There are two significant changes. First, the Routing Group Connector (RGC) uses the SMTP transport to deliver messages between routing groups. Second, it allows for the modification of the delivery time, the size, and the content, based on the designated priority (low or high), of the message.
SMTP Connector. The SMTP Connector utilizes DNS MX records for mail delivery. This differs from the Routing Group Connector in that it is point-to-point and only uses DNS for hostname resolution. It has the ability to take advantage of Windows 2000 Certificate Server and Exchange 2000 Key Management Services as a means to provide secured communications between routing groups.
x.400 Connector. The x.400 Connector was developed to be used with low-bandwidth links, in situations where there is a significant amount of delay in response time and/or where the WAN connection is metered based on activity. This is achieved by minimizing wire traffic.
The most common connector used is the Routing Group Connector. The reason this is chosen is based on ease of configuration.
Understanding Link State Tables
Earlier in this chapter, the issue of cost in its relationship to routing was mentioned. This section is designed to go into further detail with regard to Link State Tableswhere they are stored, the information they contain, and the function they serve in the Exchange 2000 environment.
The Link State Table (LST) contains the list of the routes, the status of the connectors that make up the routes, and the defined costs. It also contains a local copy of the Exchange routing topology for the entire Exchange organization. The Link State Table functions in much the same way as the Open Shortest Path First (OSPF) routing protocol. The routing group master maintains the master copy of the Link State Table that is contained in the Configuration NC.
Cost Utilization for Transfer
To better understand the cost utilization and message routing function, it is important to break it down and analyze each of its parts. The Link State Table has routing and cost information that is used with the Link State Algorithm (LSA) to determine the most cost-efficient route. This calculation is called on when the status of the link has changed. Up or down are the only two states available to a link. All servers within the routing group are responsible for reporting any status changes to any links in the routing group. The routing group master (RGM) acts as the gatekeeper and is assigned the duty of distributing any information regarding link status to all other servers within the routing group. The routing group bridgehead is responsible for replicating this information to other neighboring routing groups.
Creating Routing Groups for Efficient and Reliable Message Routing
A routing group bridgehead server is responsible for maintaining the link state routing information and is the conduit used to send messages between routing groups. Redundancy is one of the ways to deliver efficient and reliable message routing. This is achieved by defining multiple target bridgeheads. Multiple target bridgehead servers can be defined within each routing group. In the event that the first bridgehead server becomes unavailable, the Routing Group Connector or the connector that is being used would resort to other defined target(s) for message delivery.
There is a cost associated with each optional route. The cost is ranked according to its level of efficiency, meaning the higher the cost the less efficient the route. The cost associated with the use of a route controls the query order.
In the event that all target bridgeheads on the connector become unavailable, the whole connector would be marked down and other routes would be evaluated. If there are other available routes, message(s) will then be rerouted. If there are no other routes available, the message sits in the local queue until the connector comes back up. When designing the Exchange message routing topology, all scenarios should be considered and incorporated into the final design.
If multiple target bridgehead servers are defined, the routing group gains in two areas, load balancing and stability. With multiple servers defined, if one goes down, Exchange can always find and use another server for message transmission. Where configuration and use of Routing Group Connectors are concerned, administrators have the ability to control
Connection schedules
Message priority (high, normal, or low)
Message size limits