SearchExchange has written an article on the best and worst practices for email archiving in Exchange Server.  It is a detailed post and I agree with it, however it highlights some best practices that Exchange Server does not do very well on its own without additional systems in place.

PST Files

Outlook PST files are big disk space consumers and are not subject to inspection and retention by archiving systems.  Exchange Server itself cannot prevent the creation of PST files.  This requires the use of other systems such as Group Policy restrictions and file server storage policies.

For environments where PST files have been allowed to proliferate, Exchange Server offers very little in the way of tools to efficiently import the data into archives.  Fortunately there are third party email archive applications that do offer these tools for bulk processing of PST archives.

Separate Archival and DR Storage

The built in archiving available in Exchange Server 2010 does not separate active mailbox data from archive mailbox data.  Both types of data are stored in the same database, subject to the same backup and recovery process.

The best practice is to store and backup archive information on less expensive, high volume storage, and retain it separately (though it can be sent to the same offsite location as the active data).

During a recovery situation the active data can then be restored first as a priority, and archive data restored separately later when normal services have been resumed, because archive data is usually a much lower priority for restore in a DR situation.

Centralize Email Storage

Although Exchange Server environments are typically consolidated into fewer physical locations these days compared to previous years, in larger environments they are still often distributed into regional locations.

Archive information is best kept as centralized as possible, in one primary location with an additional offsite backup copy.  Again the archive features of Exchange Server 2010 limit this capability, because archive data is stored in the same location as the active mailbox.

Third party archiving systems that extract email data from Exchange Server and store it in separate databases are much more capable of the desired centralization.

Maintain Older Readable Versions of Archival Data

The Exchange Server database format has changed several times as the product has iterated through its version history.  Older versions of Exchange Server did not have database portability, which is the feature by which a foreign database can be mounted and read by a different Exchange Server than it was originally connected to.

Restoring very old Exchange data can often mean building a replica of the original Active Directory and Exchange environments in an isolated lab, a situation in which virtualization comes in very handy instead of trying to retain and rely on very old hardware platforms for this task.  Even so it is time consuming and cumbersome to retrieve the data.

On the other hand most third party archive systems store archives in a format independent of Exchange Server database developments, and that data is preserved and retrievable throughout any upgrades of the archiving product itself.

As good as Exchange Server is it is clear that adhering to best practices for email archiving will require organizations to look beyond Exchange at third party products that can integrate and offer more flexibility and reliability over time.