E-mail has become the lifeblood of enterprise communication in the 21st century, and Microsoft’s Exchange Server has positioned itself as virtually the only game in town. Sure, there are still organizations that use Lotus Notes or Novell GroupWise, but they are frequently migrating to Exchange. Having cornered the messaging market, the guys in Redmond, Washington are not sitting on their hands; they have provided us with a new version of Exchange that makes significant improvements over the previous version for the organization, its administrators, and the end-users.
The enterprise using Exchange 2003 will have no choice but to migrate to the 2010 version. Exchange 2003 will not be supported much longer, and the drawbacks of that product have become apparent in the days of large e-mail and bloated mailboxes. Those organizations that have already moved to Exchange 2007 will wish to consider which of the new features of Exchange 2010 will have the most impact on their messaging environment and will likely conclude that it’s worth the effort to move to the new product.
Changes for the Organization
Since the unification of the Exchange Directory with Active Directory in Exchange 2000, considerable changes have been implemented in the architecture of the product in an effort to keep up with colossal expansion of the use of e-mail and the size of the mailboxes. The improvements of Exchange that will have the most impact on the enterprise are designed to ensure that the messaging servers are reliable and robust.
The Database Availability Group (DAG)
Perhaps the single most significant change in the Exchange Server design is in the area of high availability. With Exchange Server 2007, Microsoft introduced continuous replication as an approach to providing redundancy of the user’s mailboxes. The continuous replication technologies came in three flavors: Local Continuous Replication (LCR), Cluster Continuous Replication (CCR), and Standby Continuous Replication (SCR). These technologies permitted organizations to avoid the cost and complexity of deploying the classic Single Copy Cluster (SCC), whose approach to failover clusters essentially duplicated the high availability approach of the previous versions of Exchange.
Exchange 2010 essentially eliminates LCR and SCC, and rolls the CCR and SCR into a single unified process known as the Database Availability Group (DAG). The DAG can contain as many as 16 replicas of a mailbox database to provide for truly distributed database access. The structure of the DAG provides the enterprise with an availability and storage model that can eliminate the need for costly Storage Area Networks (SANs), without sacrificing their reliability.
While speaking of our Mailbox Servers, we should note that the new version of Exchange has eliminated the use of Storage Groups, and the Mailbox Databases are no longer directly connected to the Server Object. This new configuration greatly supports the DAG architecture and allows for the management of Mailbox Databases at the Organization Level, instead of the Server Level as in Exchange 2007. This new design makes migration and recovery a much easier process.
In the Exchange community, a subject of frequent discussion has been the continued use of the Extensible Storage Engine (ESE) database as the backbone of the infrastructure. There were some who were hoping that Exchange 2010 would include the adoption of Microsoft’s SQL Server as the mailbox database. Those individuals might be disappointed to learn that 2010 does not provide us with a SQL database, yet, through re-engineering, Exchange 2010 has achieved a substantial improvement in the speed of Input/Output Operations. The Exchange Team has touted the improvement to approach 70%, and that speed, combined with the large bump seen when moving from Exchange Server 2003 to Exchange 2007, should satisfy the critics concerned about the continued use of the ESE database. An added plus in retaining the database structure is the relative ease of migration from previous versions to Exchange 2010, which surely would have suffered a major blow by moving to a SQL database.
Structurally, Exchange 2010 now has all data flow through the Client Access Server (CAS), regardless of whether it is coming in from a web or MAPI client. This new construct is intended to support the DAG design and the independence of the Mailbox Database from a specific Mailbox Server.
Note: This blog post is an excerpt of the Global Knowledge White Paper “Exchange Server 2010: What’s New and What’s Changed?” Download the rest of this article here.