By Jill Reese

“What are you doing about these unwanted messages?” is the most common question that OIT e-mail system administrators receive about e-mail service. OIT uses several tactics to limit spam (the electronic equivalent of junk mail), including blocking e-mail from our systems based on message profiling and marking messages so that the recipient can decide what to do with them. Studies have shown that about 70 percent of all e-mail on the Internet is spam. Without some limits on spam, it would overwhelm our e-mail systems and make them unusable.

The first level of defense against spam is at the mail forwarders, which receive all messages sent to ‘name@umd.edu’ addresses, and forward them to the recipients’ primary e-mail addresses, listed in the University Directory. The mail forwarders reject incoming spam messages based on dynamic, frequently updated local and shared lists of IP addresses sending spam, as well as telltale words in subject lines, like “viagra,” “dr@gstore,” and “st0ck.” Using these methods, the mail forwarders reject about 60 percent of the university’s incoming mail as spam.

After mail destined for OIT’s three major e-mail systems (WAM, Glue/DEANS, and Mail@umd) leaves the mail forwarders, it is evaluated by anti-spam filtering software, which scores and marks the messages and then passes them on to the individual e-mail accounts. The WAM and Glue/DEANS systems use SpamAssassin. This is an open source product that is designed to use a number of filtering approaches on both the message header and body to determine if a message is spam or not. Messages are scored and marked, then passed on to the recipient’s e-mail account, which may or may not have filters to determine where the e-mail will end up. If there are no filters, the marked messages will go into the recipient’s inbox. The Mail@umd system filters arriving e-mail through its own proprietary anti-spam engine. The messages are also scored and marked, but since spam filtering is automatically turned on for all Mail@umd account holders, spam messages are deposited into the recipient’s Junk Mail folder instead of the inbox.

While a person may see a message as spam, software programs and automated blocking methods are not as sensitive. Despite best efforts to mark spam, no software program is fully effective at marking all spam without also occasionally marking legitimate e-mail. Folders used to collect filtered junk mail should be checked regularly to remove the occasional legitimate message that gets marked as spam. Also, it is possible that an occasional legitimate message could be rejected by the mail forwarders. If a message is rejected by the forwarders, the sender of the message will get a bounce-back message. If you or one of your e-mail correspondents finds that legitimate e-mail is being rejected, please call the OIT Help Desk at 301.405.1500 to have the problem resolved quickly.

Because spammers often aim to masquerade their messages as legitimate mail and regularly change their tactics to circumvent anti-spam procedures, a 100 percent effective solution for spam prevention does not exist. Still, OIT employs several strategies to limit the amount of spam that arrives in university inboxes.

Learn how spammers get your address