[Macpartners] Fwd: Spam Screening @ MIT: SpamAssassin's software version schedule for upgrade
Ginny Williams
ginnyw at MIT.EDU
Thu Nov 20 12:33:03 EST 2003
--- begin forwarded text
X-Sieve: CMU Sieve 2.2
Cc: Financial Services <fss at mit.edu>, Theresa M Regan <tregan at mit.edu>,
Information Systems <infosys at mit.edu>
From: Theresa M Regan <tregan at MIT.EDU>
Subject: Spam Screening @ MIT: SpamAssassin's software version
schedule for upgrade
Date: Wed, 19 Nov 2003 20:40:07 -0500
To: IT-Partners <itpartners at mit.edu>
X-Mailer: Apple Mail (2.606)
X-Spam-Score: -6.3
X-Spam-Flag: NO
X-Scanned-By: MIMEDefang 2.28 (www . roaringpenguin . com / mimedefang)
Date: November 19th, 2003
To: IT-Partners
From: Information Systems
Subject: Spam Screening @ MIT: SpamAssassin's software version
schedule for upgrade
---------------
Information Systems reminds the Community that MIT Spam Screening is
a service readily available to all colleagues who receive their
e-mail via the MIT Mail System. All e-mail received by the post
office servers (po9, po10, po11, po12, po14) is scored for the
"likelihood" of unwanted junk mail, often referenced as spam mail.
The handling of e-mail scored as "spam" varies depending upon one's
e-mail client and tailored configuration.
The upcoming software upgrades are focused on the post office servers
with no direct impact on desktop spam filter settings or servers
setting. MIT colleagues who receive their e-mail via the MIT Mail
System will benefit immediately. Now, may be an opportune time to
remind your Departments and staff about MIT's Spam Screening service.
Further information is available at:
<http://web.mit.edu/is/services/email/nospam/>
Information Systems will stage the upcoming upgrade to our Spam
Screening implementation. MIT's Spam Screening package,
SpamAssassin, performs a series of tests on incoming e-mail message,
scores it according to a set of criteria indicating the probability
an e-mail message may be spam.
The upgrade will include:
* Upgrade to SpamAssassin version 2.60
* Addition of a learning Bayesian filter based on the algorithms
described at <http://www.paulgraham.com/>
and
<http://radio.weblogs.com/0101454/stories/2002/09/16/spamDetection.html>
* Ability for IMAP users to train the Bayesian filter to
correctly identify messages it misidentifies
Training the Bayesian filter with mail you received which was
not identified as Spam simply requires you to move the message to
your Spamscreen folder. Each night the post office will learn from
the messages you place there, and it will help correctly identify
these messages in the future.
* Addition of a variety of new internet blacklists for checking
e-mail relays used in sending messages
* Various bug fixes and enhancements to keep up with the
adaptations of spammers
Internal testing of this Spam Screening implementation shows an
increase in correctly identifying of unwanted e-mail by 20%. Each
Post Office server will be upgraded at 1am on the following schedule:
po12.mit.edu on Tuesday, November 25th
po14.mit.edu on Friday, November 28th
po9.mit.edu on Sunday, November 30th
po10.mit.edu on Tuesday, December 2nd
po11.mit.edu on Thursday, December 4th
MIT Services Status website, 3DOWN, <http://is3down.mit.edu/>, will
be updated during these software upgrades.
In closing, our thanks to everyone who has assisted with introducing
MIT Spam Screening to their Departments and colleagues. As we
continue to hear of the large volume of spam e-mail, daily and
estimate that 25% of the e-mail received at MIT is spam, please
assist us by ensuring that all colleagues are aware and participating
in this service. More complete information about MIT Spam Screening
is available at:
<http://web.mit.edu/is/services/email/nospam/>
--- end forwarded text
More information about the Macpartners
mailing list