Kaspersky Lab kaspersky anti-spam 2.0 マニュアル

ページ / 133
Kaspersky Anti-Spam 2.0 Enterprise Edition / ISP Edition 
9
 
with caution to avoid false triggering. 
•  absence of the sender’s server in the DNS; 
•  matching of one of the message headers a regular expression stored by 
the program; 
•  excessively large size of the message. 
 
For more details on e-mail message analysis, see para 4.3.1, page 26.  
Secondly
,  content filtering is used, i.e. the message content is analyzed 
(including the header Subject) and attached files.
1
 Linguistic algorithms based on 
a comparison with sample messages and regular expressions (words and word 
combinations) are used. 
Messages that do not match the parameters of a filtering rule or of a content 
filtering algorithm are sent to the recipients unaltered. 
Messages that display features characteristic of spam will be processed in 
accordance with the filtering rules (see para 4.3, page 25). 
Profiles (sets of filtering rules), lists of addresses these rules refer to, and other 
Kaspersky Anti-Spam settings can be modified by the mail service administrator 
using the WebConfigurator program. 
 
Kaspersky Anti-Spam is distributed with preinstalled filter profiles (sets 
of filtering rules) that ensure effective spam detection and provide for 
several options of spam processing. Before you start using the filter, 
make sure that you have reviewed the filtering schemes suggested (see 
para 4.4, page 35) and choose the scheme that suits you better. 
If you wish to modify these preinstalled filter profiles or to create new 
profiles, please carefully read chapters of this manual describing the 
filter operation logic (see para 4.3, 25) and its setup using the 
WebConfigurator program (Chapter 5, page 47). 
Pay special attention to the filter setup. An Improper setup may result 
in: 
•  inefficient filter operation (most of the unsolicited messages pass 
through the filter) 
•  the loss of normally allowed e-mails. 
                                                           
1
 Plain text, HTML, Microsoft Word and RTF formats analysis is provided. For more details 
see para 4.3.2, page 27.