Maxim_Kazmin - Fotolia

Words to go: Exchange Server architecture

Microsoft's Exchange team released its recommendations for the preferred architecture for Exchange 2013 and 2016. Here are some terms Exchange architects should know.

This glossary will help you keep track of some of the terminology you'll encounter while planning your Exchange Server architecture.  

Preferred Exchange architecture is the Exchange Engineering team's best-practice recommendation for deploying Exchange Server 2013 and 2016. Preferred architecture is divided four design areas: namespace, data center, server and database availability group. Not all of the terms Exchange architects will encounter while planning an Exchange Server architecture are commonly discussed in the course of daily Exchange management, or completely understood in the context of an architecture for deployment.

Bookmark this page or print out a copy to keep on your desk for a quick reference.

Bound namespace: In a bound namespace, a different namespace is used for each data center in a site-resilient pair. There are two database availability groups (DAGs) deployed in the data center pair, and each DAG contains a set of mailbox databases for a particular data center. The DAG is configured in an active-passive topology. In the bound namespace model, primary and failback namespaces are preferred to prevent clients from trying to connect to a data center where they have no connectivity.

Preferred Exchange architecture is the Exchange Engineering team's best-practice recommendation for deploying Exchange Server 2013 and 2016.

Unbound namespace: In an unbound namespace, there is only one database availability group deployed across a data center pair. The DAG has mailbox servers in each data center. The same namespace is used across both data centers in a site-resilient pair, and the DAG is configured in an active-active topology. In the unbound namespace model, a single namespace is preferred, because each of the data centers in the pair can service the user request.  

Load balancing: Load balancing provides failover redundancy in case of a computer failure. It also enables an Exchange Server deployment to handle more traffic than one server can process while offering a single host name for clients.

Session affinity: Session affinity means that once a request comes into a site from a client, all further requests will be routed to the same server that served the first request. Session affinity is also referred to as sticky session.  

Multirole physical server: In Exchange architecture, multirole physical servers are deployed instead of virtualized hardware, because all of the servers have the same hardware, configuration options and installation process. The multirole servers distribute client access and mailbox resources across a larger pool of servers.

Active Directory site: An Active Directory site represents the topology or physical structure of a network. A site is a set of computers that are well-connected by a high-speed network, such as a local area network.  

Domain name system round robin: DNS round robin provides a client connection with the IP address of each mailbox server. The first client connection request receives the first IP address, and subsequent requests receive the next address in the list.

Next Steps

Exchange Server 2016 architecture planning

The importance of geo load balancing in Exchange 2013

How to configure a geo load balancer in Exchange 2013

Dig Deeper on Exchange Server setup and troubleshooting