Campus-wide Spam Filtering
The campus e-mail system has become increasingly
strained by the influx of more and more Spam every day. To
combat this, the University has decided to implement a Spam filtration
system at the campus-wide level. This system will employ a
collection of tests to ascertain if a message is Spam. Each
message will undergo this testing and be assigned a numerical "Spam score". Messages that are determined to be
blatant Spam
will be removed from the campus e-mail system. All other messages
will be marked with the Spam score and delivered.
This system will provide two benefits to the campus community. First, the messages that are most blatantly Spam will be removed and will no longer impact resources at any level. Second, the ratings on the potential Spam messages will allow for the individual users to set up rules on their e-mail client to remove messages they feel are Spam based on their individual tolerance for Spam-like e-mail.
This two-level system minimizes the potential problems from false-positives. False-positives are e-mail messages that the user wishes to receive, but are assigned a high enough Spam score that a low tolerance setting would delete the message. By setting the campus-wide tolerance level sufficiently high, we can ensure that virtually no legitimate e-mail messages are being deleted. This rating system is sufficiently advanced that it will allow a reduction the total Spam delivered by a significant percentage, even with a reasonably high setting that will avoid false positives.
The system will be implemented using a commercial product that operates at the main connection between the Internet and campus e-mail systems. Any e-mail bound for a campus address will pass through this machine and be analyzed before entering the campus e-mail systems. Thus, every campus e-mail account will benefit from this protection.
The goal of this implementation is to be as transparent to the user as possible. The removal of blatant Spam will be done silently, so the user need not wade through hundreds of messages a week informing them that Spam has been removed -- indeed, that would simply become yet another type of Spam on its own. The Spam scores attached to each e-mail message delivered will not be visible in the body of the message. If you wish to view the Spam rating you must look at the message headers. Finally, the rules on the user's e-mail client can be used to further eliminate Spam. The overall effect for the end-user will be a significant reduction in the amount of Spam received.
This system will provide two benefits to the campus community. First, the messages that are most blatantly Spam will be removed and will no longer impact resources at any level. Second, the ratings on the potential Spam messages will allow for the individual users to set up rules on their e-mail client to remove messages they feel are Spam based on their individual tolerance for Spam-like e-mail.
This two-level system minimizes the potential problems from false-positives. False-positives are e-mail messages that the user wishes to receive, but are assigned a high enough Spam score that a low tolerance setting would delete the message. By setting the campus-wide tolerance level sufficiently high, we can ensure that virtually no legitimate e-mail messages are being deleted. This rating system is sufficiently advanced that it will allow a reduction the total Spam delivered by a significant percentage, even with a reasonably high setting that will avoid false positives.
The system will be implemented using a commercial product that operates at the main connection between the Internet and campus e-mail systems. Any e-mail bound for a campus address will pass through this machine and be analyzed before entering the campus e-mail systems. Thus, every campus e-mail account will benefit from this protection.
The goal of this implementation is to be as transparent to the user as possible. The removal of blatant Spam will be done silently, so the user need not wade through hundreds of messages a week informing them that Spam has been removed -- indeed, that would simply become yet another type of Spam on its own. The Spam scores attached to each e-mail message delivered will not be visible in the body of the message. If you wish to view the Spam rating you must look at the message headers. Finally, the rules on the user's e-mail client can be used to further eliminate Spam. The overall effect for the end-user will be a significant reduction in the amount of Spam received.
Last Updated: February 21, 2006