IPS configuration options
Besides configuring an IPS filter or selecting IPS signatures for an IPS sensor, you can configure additional IPS options for each sensor or globally for all sensors. This topic introduces the following available configuration options:
- Malicious URL database for drive-by exploits detection
- IPS signature rate count threshold
- Botnet C&C
- Hardware acceleration for flow-based security profiles (NTurbo and IPSA)
- Extended IPS database
- IPS engine-count
- Industrial signature database
- Fail-open
- IPS buffer size
- Session count accuracy
- Protocol decoders
Malicious URL database for drive-by exploits detection
This feature uses a local malicious URL database on the FortiGate to assist in detection of drive-by exploits, such as adware that allows automatic downloading of a malicious file when a page loads without the user's detection. The database contains all malicious URLs active in the last one month, and all drive-by exploit URLs active in the last three months. The number of URLs controlled are in the one million range.
This feature can be enabled from a IPS Sensor in the GUI by going to Security Profiles > Intrusion Prevention and editing or creating an IPS Sensor. Then enable Block malicious URLs.
From the CLI:
config ips sensor
edit <profile>
set block-malicious-url [enable | disable]
next
end
|
Blocking malicious URLs is not supported on some FortiGate models, such as FortiGate 51E, 50E, or 30E. |
IPS signature rate count threshold
You can use the IPS signature rate-based settings to specify a rate count threshold that must be met before the signature is triggered. A rate count threshold provides a more controlled recording of attack activity. For example, if multiple login attempts produce a failed result over a short period of time, then an alert would be sent and traffic might be blocked, which is a more manageable response than sending an alert every time a login fails.
This can be configured from the GUI by going to Security Profiles > Intrusion Prevention. Create or edit an IPS sensor. Within the sensor, edit the IPS signatures and filters. Only IPS signatures have the rate-based settings option. IPS filters do not.
Some settings are only available from CLI.
The syntax for this configuration is as follows:
config ips sensor
edit default
config entries
edit <Filter ID number>
set rule <*id>
set rate-count <integer between 1 - 65535>
set rate-duration <integer between 1 - 65535>
The value of the rate-duration is an integer for the time in seconds.
set rate-mode <continuous | periodical>
The rate-mode refers to how the count threshold is met.
If the setting is “continuous”, and the action is set to block, the action is engaged as soon as the rate-count is reached. For example, if the count is 10, the traffic would be blocked as soon as the signature is triggered 10 times.
If the setting is “periodical”, the FortiGate allows up to the value of the rate-count incidents where the signature is triggered during the rate-duration. For example, if the rate count is 100 and the duration is 60, the signature would need to be triggered 100 times in 60 seconds for the action to be engaged.
set rate-track <dest-ip | dhcp-client-mac | dns-domain | none | src-ip>
This setting allows the tracking of one of the protocol fields within the packet.
Botnet C&C
See IPS with botnet C&C IP blocking.
Hardware acceleration for flow-based security profiles (NTurbo and IPSA)
Some FortiGate models support a feature call NTurbo that can offload flow-based firewall sessions to network processors. See also Hardware Acceleration > NTurbo offloads flow-based processing. For IPSA enhanced pattern matching, see Hardware Acceleration > IPSA offloads flow-based advanced pattern matching.
Some FortiGate models also support offloading enhanced pattern matching for flow-based security profiles to CP8 or CP9 content processors. You can use the following command to configure NTurbo and IPSA:
config ips global
set np-accel-mode {none | basic}
set cp-accel-mode {none | basic | advanced}
end
If the np-accel-mode option is available, your FortiGate supports NTurbo. The none option disables NTurbo, and basic (the default) enables NTurbo.
If the cp-accel-mode option is available, your FortiGate supports IPSA. The none option disables IPSA, and basic enables basic IPSA, and advanced enables enhanced IPSA, which can offload more types of pattern matching than basic IPSA. The advanced option is only available on FortiGate models with two or more CP8 processors, or one or more CP9 processors.
Extended IPS database
Some models have access to an extended IPS Database. Because the extended database may affect FortiGate performance, the extended database package may be disabled by default on some models, such as desktop models.
You can only enable the extended IPS database by using the CLI.
config ips global
set database extended
end
FortiGate models with the CP9 SPU receive the IPS full extended database, and the other physical FortiGate models receive a slim version of the extended database. The slim-extended database is a smaller version of the full extended database that contains top active IPS signatures. It is designed for customers who prefer performance.
|
|
Customers with non-CP9 SPU models need to upgrade to a CP9 SPU model (physical FortiGate) in order to get full IPS signature coverage. All FortiGate models 200 (E and F) and higher have a CP9 SPU. See Determining the content processor in your FortiGate unit in the FortiOS Hardware Acceleration Guide to check if your device has a CP9 SPU. |
IPS engine-count
FortiGate units with multiple processors can run one or more IPS engine concurrently. The engine-count CLI command allows you to specify how many IPS engines to use at the same time:
config ips global
set engine-count <int>
end
The recommended and default setting is 0, which allows the FortiGate unit to determine the optimum number of IPS engines.
Industrial signature database
Industrial signatures are defined to protect Industrial Control Systems (ICS), Operational Technology (OT) and SCADA systems, which are critical infrastructure used by manufacturing industries. These signatures are enabled by default, but can be configured by using the following CLI:
config ips global
set exclude-signatures {none* | industrial}
end
Fail-open
A fail-open scenario is triggered when IPS raw socket buffer is full. Therefore IPS engine has no space in memory to create more sessions and needs to decide whether to drop the sessions or bypass the sessions without inspection.
config ips global
set fail-open {enable | disable}
end
The default setting is disable, so sessions are dropped by IPS engine when the system enters fail-open mode.
When enabled, the IPS engine fails open, and it affects all protocols inspected by FortiOS IPS protocol decoders, including but not limited to HTTP, HTTPS, FTP, SMTP, POP3, IMAP, and so on. When the IPS engine fails open, traffic continues to flow without IPS scanning.
|
|
Sessions offloaded to Nturbo do not support fail-open. When Nturbo data path is overloaded, traffic is dropped regardless of fail-open setting. |
IPS buffer size
If system enters fail-open mode frequently, it is possible to increase the IPS socket buffer size to allow more data buffering, which reduces the chances of overloading the IPS engine. You can set the size of the IPS buffer.
config ips global
set socket-size <int>
end
The default socket size and maximum configurable value varies by model. In short, socket-size determines how much data the kernel passes to the IPS engine each time the engine samples packets.
Take caution when modifying the default value. If the socket-size is too large, the higher memory used by the IPS engine may cause the system to enter conserve mode more frequently. If set too low, the system may enter IPS fail-open mode too frequently.
Session count accuracy
The IPS engine can track the number of open session in two ways. An accurate count uses more resources than a less accurate heuristic count.
config ips global
set session-limit-mode {accurate | heuristic}
end
The default is heuristic.
Protocol decoders
The FortiGate Intrusion Prevention system uses protocol decoders to identify the abnormal traffic patterns that do not meet the protocol requirements and standards. For example, the HTTP decoder monitors traffic to identify any HTTP packets that do not meet the HTTP protocol standards.
To change the ports a decoder examines, you must use the CLI. In this example, the ports examined by the DNS decoder are changed from the default 53 to 100, 200, and 300.
config ips decoder dns_decoder
config parameter "port_list"
set value "100,200,300"
end
end
You cannot assign specific ports to decoders that are set to auto by default. These decoders can detect their traffic on any port. Specifying individual ports is not necessary.