ICUK Dedicated Server Solutions
 
 
 
" I would like to commend you guys on your support. You always answer promptly and with an answer to the actual question! Our previous hosts were terrible at support - we didn't get any!

I will be swapping my own personal hosting to you guys in the future because of the excellent support!
"
Sandra

Spam & Virus Filtering

Our Spam + Virus Filtering product is a result of several man hours of testing, configuration, research and development by our in house staff allowing to use your email without hindrence from junk and virus bourne mail, however formed and sent to you.

The Filter works on a separate and extremely powerful mail server cluster to our regular POP3 mail delivery machines; preventing a deluge of spam from impacting on normal mail delivery. Each email is scanned and checked against thousands of different rules to determine the likelyhood of it being junk or geniune mail.

Heuristics Engine
Thousands of tests check the message headers and content for tell tale signs of spam. The originating mail server and senders address are verified to prevent spoofing. We proactively monitor and adapt the tests used by by the filter to any new techniques the spammers employ.

Bayesian Probability Algorithms
Bayesian algorithms are used to build up a database of the characteristics of unwanted mail and the characteristics of legitimate mail. The database can then be consulted to determine the likelihood of any particular message being spam. Furthermore, through continually analysing messages the database is actively updated as the nature of spam changes, creating a form of self evolving intelligence.

Checksum Databases
Any message has a unique checksum, as users report unsolicited mail messages, their checksums are added to a database. Several collaborative checksum databases are consulted to ascertain whether the message concerned matches one previously identified as spam by another Internet user.

Real Time Blackhole Lists
Several real time blackhole lists (RBLs) are maintained on the Internet which catalog known sources of spam. The Spam filter checks for the presence of the originating mail servers in any of these lists.

Vipul's Razor
As well as the standard networking tests, the new version will utilise a "Razor client" to link in with Vipul's Razor network. This technology has not been used before now due to the large overhead that is involved in its use.

The following summary of Vipul's Razor, taken directly from the author, explains the ins and outs of the protocol:

Vipul's Razor is a distributed, collaborative, spam detection and filtering network. Through user contribution, Razor establishes a distributed and constantly updating catalogue of spam in propagation that is consulted by email clients to filter out known spam. Detection is done with statistical and randomised signatures that efficiently spotmutating spam content. User input is validated through reputation assignments based on consensus on report and revoke assertions which in turn is used for computing confidence values associated with individual signatures.

Pyzor
Pyzor is a Python implementation of Razor, which runs much in the same way but via an alternative network. Consulting both networks greatly improves the filter's capability to catch otherwise low scoring messages.

OCR
The introduction of Optical Character Recognition scanning techniques is the latest addition that we are most excited about. With the latest trend of spam messages bestowing sole image files, previous attempts atfiltering them have largely been ineffective and has affected many spamfiltering providers.

With this new method, as well as the body and headers of messages being scanned for common spam words and phrases, any attached or inline images are also scanned using OCR functions.

Whilst some images are crafted in such a way that they simply cannot be recognised by OCR, we have seen that over 75% of messages containing spam images are detected.

Additional Rules
As well as the above new methods, new filtering rules have been added alongside the existing rules to tackle common spam themes such as:

* Adult orientated
* Message obfuscation
* Stock/finance related
* Spoofed legitimate emails (including phishing techniques)
* OEM software
* Drug and performance related medicines

Virus Scanner
Virus definitions are checked for and updated automatically ensuring that you are protected against threats as they emerge. The virus scanner is capable of detecting all known virus types, including polymorphic or self-encrypting viruses, it is even capable of scanning the contents of compressed archives such as zip and rar files ensuring that no virus bearing email can make it to your inbox.

Detailed Reports
Each message scanned and found to be spam can include a detailed text attachment explaining why the message was considered as such, including the points awarded to each rule that was matched and the total points scored on the email.