Previous Section  < Day Day Up >  Next Section

7.2 The Rule Sets

Here is a brief description of the rule sets:


My personal favorite set of rules. They detect when a host on your local network is sending a known response to a successful attack. While it might not be as useful as catching the attacker before he has succeeded, the alerts these rules generate are very often not false positives. There are some that are a little noisy—in particular the rule that alerts on a "403 - Forbidden" HTTP response.


Detects traffic generated by backdoor network connections, including those created by attackers using many rootkits and stealthy remote control applications (like subseven, netbus, and deepthroat).


Watches for illegal packet header settings like a TCP and UDP port 0 traffic, or a SYN packet to a multicast address.


Disabled by default. It watches for people using instant messengers and other Internet chat protocols. If this activity is against your organization's security policy, enable this rule set.


Alerts on traffic generated by many well-known distributed denial-of-service mechanisms, including Trin00 and shaft. The Stacheldraht rules can be noisy, since they are just looking for specific words in the payload that may be common in your environment.


Actually not referenced by default in the snort.conf file; really just a museum of old Snort rules.


Alerts on attacks against DNS servers (including detection of zone transfers).


Detects traffic generated by known denial-of-service attacks. It will detect some specifically named attacks like winnuke and jolt, but will also detect classes of attacks like IGMP and teardrop attacks.


This is where new types of rules are included. This file is included by default, so check it for new rules. Very often this is just an empty file.


Includes the signatures of many known exploits. Seeing an alert generated by these rules should not cause immediate panic. They indicate that an exploit attempt has occurred. When you see one of these alerts, verify that the target system is, in fact, vulnerable to the attack. Hopefully, you have the system patched or updated to address the vulnerability the exploit attempted to attack.


Alerts on many known types of attack against the finger service that runs by default on many Unix-based operating systems. If you have the finger service disabled on your systems (I do, where possible), you can disable this rule set.


Generates alerts when known attacks are detected against the FTP service.


This noisy little set of rules is disabled by default. It may be useful when troubleshooting a specific ICMP problem on your network, but in general it just generates noise and should remain disabled.


Alerts when it sees the signs of pings specific to particular attack tools. These alerts can be very useful. The "destination unreachable" rules can be very noisy, though, and if you choose to keep this set of rules enabled, consider turning off the loud "unreachable" rules.


Generates alerts when known attacks against the IMAP email service are detected.


Disabled by default. They generate alerts on a variety of traffic that is normally found on a healthy, secure network. They may be useful in troubleshooting some issues, though.


Stores rules that you create.


Contains rules that don't fit easily into other categories. In my experience, they generate a large number of false positives on traffic that should not be a concern on a network that utilizes reasonable defense-in-depth strategies. I usually disable this entire rule set.


Disabled by default. If it is against your organization's security policies to run multimedia applications across the network, enable this rule set.


Detects known attacks against mySQL database servers.


Detects several of the recent Windows-attacking worms that are causing headaches for network and system administrators around the world. Some of the alerts can generate false positives (namely the administrative share access rules and the rules that alert on SMB and NetBIOS access). If the Snort sensor is watching Internet traffic only and NetBIOS traffic is not allowed in or out of your environment, consider disabling this rule set.


Contains signatures that indicate attack against network time protocol servers.


Detects known attacks against Oracle database servers.


Watches for traffic generated by other IDSs. If you are the only person authorized to run an IDS in your environment, this can be a relevant concern. You can likely disable this rule set.


Disabled by default. They detect activity generated by peer to peer software. Peer to peer clients can represent a dangerous vector for the introduction of viruses, worms, and other malicious code into your environment. If your organization has policies against the use of such software, enable this rule set.


Disabled by default. It contains rules that watch for activity that may against some organization's security policies (for example, an alert will be generated by PC Anywhere and VNC traffic, or an anonymous FTP login). Consider reviewing the rules in this file and leaving those that match your policies enabled, while disabling the rest (after enabling the rule set itself, of course).


Generates alerts when known POP2 email service attacks are detected.


Generates alerts when known POP3 email service attacks are detected.


Disabled by default, these rules will alert when a variety of off-color packet contents go by on the wire. If it is against your organization's security policy to visit such content on the Internet, consider enabling the rule set (but be prepared for some... interesting alerts).


Generates alerts on attacks against the remote procedure call (RPC) services employed by nearly every operating system. If the sensor is watching Internet traffic and RPC activity is not allowed in or out of the environment, consider disabling this set of rules. If the sensor is watching internal traffic, some tuning may be necessary, depending on the types of systems running on the network.


Alerts when rservices (rlogin, rsh, and rexec) are detected on the network. These are powerful commands to control remote systems. If you use these in your environment, you may want to disable this rule set.


Detects a variety of different network and service scans, from deceptive portscans to SSH and UpnP service scans. It detects the signatures of some specific scanning tools, too.


Disabled by default. It will detect shellcode in the packet payload that is attempting to compromise a variety of systems. This shellcode may be the payload of a successful attack that does not have its own signature. Since these rules are designed the check the payloads of all traffic, they can cause a significant performance hit when enabled.


Generates alerts when known SMTP email service attacks are detected.


Detects a variety of SNMP traffic. SNMP is used to manage devices on a network and many vulnerabilities have been detected in the protocol. If you are running a sensor that is only watching Internet traffic and SNMP traffic is not allowed in or out, you can disable this rule set.


Detects known attacks against Microsoft SQL Server database servers.


Alerts on dangerous traffic transmitted in telnet sessions.


Alerts on attacks against the TFTP service.


Disabled by default. This rule set is not being actively maintained and the rules really just watch for a variety of file extensions transmitted in email traffic. The real virus signatures are located within the specific service's rule sets now.


Disabled by default. It generates alerts when known generic attacks against web servers are detected. Consider enabling it, since it does not generate a large number of false positives.


Generates alerts when known attacks against CGI services are detected.


Generates alerts when potentially dangerous web client traffic is detected. Most of these alerts are based on Microsoft Outlook Web Access traffic and generate a large number of false positives. Consider disabling this rule set.


Generates alerts when known attacks against Coldfusion web application services are detected.


Generates alerts when known attacks against Frontpage web authoring services are detected.


Detects known attacks against Microsoft Internet Information Server (IIS) web servers.


Generic web attack detection rules.


Detects attacks against web servers running PHP applications (primarily runs on Apache, but it is possible to run on IIS).


Detects attacks against remote X-Windows sessions.

    Previous Section  < Day Day Up >  Next Section