Virtual servers are configured and inherit the overall configuration. Separate domains could be managed by creating multiple virtual servers, although the effect would be clumsy. Filtering can go to the level of individual addresses, with regular expression-based filtering on any mail header. The users themselves have little control over the process, which is disappointing.
RDNS checks are done on mail sources and RBL support is included (without it, the spam filtering is not particularly impressive).
Handling options are basic. Junk mail can be deleted, forwarded, or tagged with a header. The product does not use spam signatures and so does not receive updates, which limits filtering.
Advanced spammers will have little difficulty beating IMail filters, but the majority of the spam in our test spool has readily identifiable hallmarks. Even without advanced filtering, some basic rules helped IMail catch a healthy amount of spam with no false positives.
Logging is basic too: each entry has a time stamp, a message ID and a spam score, but no explanation of how the score was calculated (making administration tricky). You have the option of logging to a local text file or a log server.
The PDF documentation is adequate, but the context help is better. Once in the GUI, we had no need for the user manual. IMail is easy to set up and maintain, but should be behind a more accomplished anti-spam product. You do not want your primary messaging server doing front-line filtering.