Balancing reasonable mailbox size with acceptable backup and restore times is a continuing challenge in the Exchange world. UPDATE reader and Exchange consultant Nathan Black shared the solution he has developed, along with some insights into different types of Outlook users.
Black has identified three types of email users. Type A users delete everything fairly quickly and never hit their mailbox-size ceiling, so administrators generally don't need to worry much about them. The two other types store messages for future use, either because the messages contain some business value (Type B) or because they answer a question that will probably come up again (Type C). Type B users are fairly organized—they use some kind of folder structure to make it easy to locate items, filing messages by project name, for example. Their mailboxes might grow steadily, but they seldom contain nonbusiness mail. Type C users are disorganized. Their Inboxes may have 1000 or more items, including daily newsletters from a year or two ago, and their mailboxes grow rapidly. Fortunately, says Black, Type C users appear to comprise only about 10 percent of the population.
Black's solution, which he calls "Home Folders," is aimed mainly at the Type B user-–one with a steadily growing mailbox but an organized practice for deleting and refiling messages-–but might also help Type C users get their mailboxes under control. He sets up primary mailbox servers with 50GB to 70GB mailbox stores and enforces a mailbox limit of 70MB. However, Black doesn't encourage users to export old but still useful information to personal folder .pst files. Instead, a set of public folder servers with larger (100GB to 150GB) stores provides each Type B and C user with his or her own "home folder," in which the users can create whatever hierarchy they need to store their older items. Users have full responsibility for moving items into their home folders but must move items manually. (Outlook's AutoArchive feature can use only a personal folder .pst file, not a public folder, as its target.) The home folder servers have Deleted Item Retention turned on so that users can recover items that they inadvertently remove from their home folders.
Restoring one of the mailbox servers takes about an hour, and the "archive" public folder servers have recovery times of 3 to 5 hours. Black cites two main drawbacks of this home folders approach. One is that public folders let users search only one folder at a time, unless you include an index server in the configuration. The other is that the email portion of the organization's document-retention/destruction plan rests on users, so education and supervision are crucial.
Black isn't the only one to come up with an archive solution that relies on Exchange as the data repository. The same approach is at the heart of C2C Systems' Archive One ( http://www.c2c.com ), which addresses some of the user issues that Black identified. Archive One moves items automatically from primary mailbox storage to a long-term-storage Exchange server, based on policies that administrators establish. The archiving process has an option to leave behind in the original folder a small (less than 1KB) "ghost message" that links to the archived item. Thus, the fact that items have been archived can be essentially transparent to the user.
C2C General Manager Dave Hunt says that using Exchange as the sole data store means that administrators don't need to use different tool sets, one for Exchange management and another for managing the archive database. He suggests that organizations using Exchange 2000 Server might want to use Microsoft SharePoint Portal Server (formerly code-named Tahoe) to index the contents of both front-end mail and back-end archive servers, thus working around the subfolder-search limitation that Black raised.