Reputation preprocessor provides basic IP blacklist/whitelist capabilities, to block/drop/pass traffic from IP addresses listed. In the past, we use standard Snort rules to implement Reputation-based IP blocking. This preprocessor will address the performance issue and make the IP reputation management easier. Reputation preprocessor runs before other preprocessors.
Sections: Configuration IP list file format Events Shared memory support
The preprocessor configuration name is “reputation”.
Option Argument Required Default
priority [blacklist whitelist] No priority whitelist
nested_ip [inner outer both] No nested_ip inner
white [black trust] No white unblack
memcap = 1 - 4095 Mbytes
preprocessor reputation:\ < memcap number >,\ < scan_local >, \ < blacklist < list filename »,\ < whitelist < list filename »,\ < priority [blacklist whitelist] >,\ < nested_ip [inner outer both] >,\ < white [unblack trust] > Options:
< memcap number >:
maximum total memory allocated (in Megabytes). It can be set up to
< scan_local > : Enable to inspect local address defined in RFC 1918: 10.0.0.0 - 10.255.255.255 (10/8 prefix) 172.16.0.0 - 172.31.255.255 (172.16/12 prefix) 192.168.0.0 - 192.168.255.255 (192.168/16 prefix)
< list filename > : The IP lists are loaded from external files. It supports relative paths for inclusion and $variables for path. Multiple blacklists or whitelists are supported.
Note: if the same IP is redefined later, it will overwrite the previous one. In other words, IP lists always favors the last file or entry processed.
< priority > :
Specify either blacklist or whitelist has higher priority when source/destination is on blacklist while destination/source is on whitelist. By default, whitelist has higher priority. In other words, the packet will be passed when either source or destination is whitelisted.
Note: this only defines priority when there is a decision conflict, during run-time. During initialization time, if the same IP address is defined in whitelist and blacklist, whoever the last one defined will be the final one. Priority does not work on this case.
< nested_ip >: Specify which IP address to be used when there is IP encapsulation.
< white >: Specify the meaning of whitelist. When white means unblack, it unblacks IPs that are in blacklists; when white means trust, the packet gets bypassed, without further detection by snort. You can only specify either unblack or trust.
Note: when white means unblack, whitelist always has higher priority than blacklist.
Default configuration # Standard blacklisting. # If running in inline mode the packets will be dropped. preprocessor reputation:\ blacklist /etc/snort/default.blacklist, \ whitelist /etc/snort/default.whitelist
Inspect inner and outer IP configuration
# Inspect both inner and outer, Also adds a whitelist entry
# to make exceptions.
preprocessor reputation: n
nested_ip both, \
blacklist /etc/snort/default.blacklist, \
# Blacklisting with scan local network, use both headers,
# and whitelist has higher priority.
# Also adds a whitelist entry to make exceptions. preprocessor reputation: \ memcap 200, scan_local, nested_ip both, \ priority whitelist, \ blacklist /etc/snort/default.blacklist, \ whitelist /etc/snort/default.whitelist, \ white trust
Variable path/file configuration
$REP_BLACK_FILE1 = ../dshield.list $REP_BLACK_FILE2 = ../snort.org.list preprocessor reputation: \ blacklist $REP_BLACK_FILE1,\ blacklist $REP_BLACK_FILE2
IP List File Format
Syntax The IP list file has 1 entry per line. The entry can be either IP entry or comment.
IP List File Example ———————- # This is a full line comment 172.16.42.32/32 # This is an inline comment, line with single CIDR block
A user wants to protect his/her network from unwanted/unknown IPs, only allowing some trusted IPs. Here is the configuration:
preprocessor reputation: \ blacklist /etc/snort/default.blacklist whitelist /etc/snort/default.whitelist
In file “default.blacklist”
# These two entries will match all ipv4 addresses 220.127.116.11/1 18.104.22.168/1
In file “default.whitelist” 22.214.171.124 # sourcefire.com 126.96.36.199 # google.com
The preprocessor uses GID 136 to register events.
1 Packets are blacklisted. 2 Packets are whitelisted. 3 Packets are inspected.
No rule options
In order to minimize memory consumption when multiple Snort instances are running concurrently, we introduce the support of shared memory. After configured, all the snort instances share the same IP tables in shared memory.
System requirement This feature is supported only in Linux.
A new option, –enable-shared-rep is introduced to ./configure command. This option enables the support for shared memory. In order to signal Snort to reload the IP file lists, control socket feature should be enabled also. –enable-control-socket
If the build supports shared memory, this configuration will enable shared memory. If this option isn’t set, standard memory is used. This option must specify a path or directory where IP lists will be loaded in shared memory. One snort instance will create and maintain the shared IP lists. We use instance ID 1, specified in the snort -G option to be the master snort. All the other snort instances are clients (readers).
This option changes the period of checking new shared memory segment, in the unit of second. By default, the refresh rate is 60 seconds.
Steps to configure shared memory
1) When building Snort, add option -enable-shared-rep and –enable-control-socket to ./configure. For example: ./configure –enable-gre –enable-sourcefire –enable-flexresp3 –enable-pthread –enable-linux-smp-stats –enable-mpls –enable-targetbased –enable-shared-rep –enable-control-socket
2) Put your IP list file into a directory, where snort has full access For example: /user/reputation/iplists In order to separate whitelist with blacklist, you need to specify whitelist with .wlf extension and blacklist with .blf extension. 3) In snort config file, specify shared memory support with the path to IP files. Example: shared_mem /user/reputation/iplists If you want to change the period of checking new IP lists, add refresh period. Example: shared_refresh 300 4) Start shared memory master(writer) with -G 0 option. Note: only one master should be enabled. 5) Start shared memory clients (readers) with -G 1 or other IDs. Note: for one ID, only one snort instance should be enabled. 6) You will see the IP lists got loaded and shared across snort instances!
Reload IP list using control socket
1) Run snort using command line with option –cs-dir
Using manifest file to manage loading (optional)
Using manifest file, you can control the file loading sequence, action taken, and support zone based detection. You can create a manifest file named “zone.info” in the IP list directory.
When Snort is signaled to load new lists, a manifest file is read first to determine which zones the IPs in each list are applicable to and what action to take per list (Block, White, Monitor).
Files listed in manifest are loaded from top to bottom. You should put files that have higher priority first. In manifest file, you can put up to 255 files. Without manifest file, files will be loaded in alphabet order.
Here’s the format of the manifest file. Each line of the file has the following format:
<filename>, <list id>,<action>[, <zone>]+ <list id> ::= 32 bit integer <action> ::= "monitor"|"block"|"white" <zone> ::= [0-1051]
Using manifest file, you can specify a new action called “monitor”, which indicates a packet needs to be monitored, but does not disable detection. This is different from “block” action, which disables further detection. This new action helps users evaluate their IP lists before applying it.
An example manifest file:
#ipreputation manifest file white.wlf, 111 ,white, black1.blf, 1112, black, 3, 12 black2.blf, 1113, black, 3, 12 monitor.blf,2222, monitor, 0, 2, 8