Spam has long been a serious problem with all e-mail users. However, with recent estimates stating up-to 80% of all e-mail sent is Spam, there have been many technologies developed to help combat it.
The most common one is SpamAssassin, which is set-up with all our servers by default. However, this alone is not enough. SpamAssassin simple reads each e-mail as it's accepted by the server and through complex statistical analysis, decided whether or not it's Spam.
The following systems assist SpamAssassin in deciding what is and is not Spam, while the last one works with the mail software to work out if it should accept the mail from the server requesting to send it.
Rules-du-Jour
Although SpamAssassin performs complex statistical analysis, it needs a set of rules to work with first so it knows what it is looking for. Normally, these rules come as part of the package, and as such, only get updated with each new release.
The Rules-du-Jour program is a simple script which can be run each night to download a new set of extra rules. This helps SpamAssassin to keep up-to-date with the latest phrases and bypass attempts circulating and hence gives it a much better chance of catching newer forms of Spam.
DCC
DCC stands for Distributed Checksum Clearhosing and works by fingerprinting each e-mail. Each e-mail that is read and processed is fingerprinted by calculating it's checksum. This is then passed onto the one of the many DCC servers available on the internet.
By monitoring the traffic of e-mails across millions of users, common fingerprints of Spam e-mails can be found and then blocked (or at least in this case, marked as Spam). While this does not catch Spam in it's very early stages, it is very good at marking out spam quickly and can be very reliable in the long-term.
Spam which changes a lot or it not very common will not be picked up via this method – this is where the statistical analysis from the available rules works better.
More information can be found here.
MAPS
MAPS is another Spam prevention technology and stands for Mail Abuse Prevention System.
However, it doesn't work on the e-mail, but on the server. With each new connection from another server wishing to send mail, our servers query a number of centralised databases that contain lists of addresses that are known to either send e-mail (they have been reported and confirmed), or from a computer which may have been compromised or have been poorly set-up (for example, as an open relay) and therefore may be able to send Spam.
By looking up this database, our server can decide if this e-mail if from a genuine source or a possible spam source, and, if the latter, refuse to connect and accept the e-mail.
This helps to reduce the load on the server and initial tests have show it can reduce Spam e-mail being received by at around 50%, depending on the Spam and the user.
There is a trade-off with this setting. If you are on an address that has been listed, or someone is trying to send you a mail via a server that is listed, then our server will not allow it to be received. We have carefully chosen the databases we have used, so this should be a very rare case.
