Cisco Traffic Anomaly Detector Module Configuration Guide (Software Version 6.0)
Configuring Policy Templates and Policies

Table Of Contents

Configuring Policy Templates and Policies

Understanding Zone Policies

Understanding Policy Templates

Understanding the Different Policy Template Types

Configuring Policy Template Parameters

Configuring the Maximum Number of Services

Configuring the Minimum Threshold

Configuring Policy Template States

Configuring All Policy Template Parameters Simultaneously

Understanding the Policy Path

Understanding and Managing the Policy Services

Adding a Service

Deleting a Service

Understanding the Packet Types that the Detector module Monitors

Understanding the Traffic Characteristics that the Detector Module Monitors

Configuring Policy Parameters

Changing the Policy State

Configuring the Policy Threshold

Setting the Policy Threshold

Setting the Threshold as Fixed

Configuring a Threshold Multiplier

Multiplying a Threshold by a Factor

Configuring Specific IP Thresholds

Configuring the Policy Timeout

Configuring the Policy Action

Configuring the Policy Interactive Status

Understanding Worm Policies

Configuring Worm Policies

Identifying Worm Attacks

Monitoring Policies

Displaying Policies

Displaying Policy Statistics

Backing Up the Policy Configuration


Configuring Policy Templates and Policies


This chapter describes the Cisco Traffic Anomaly Detector (Detector module) zone policies, policy structure, and policy templates, and it describes how to configure the zone policy and the policy template parameters.

This chapter refers to the Cisco Guard (Guard), the companion product of the Detector module. The Guard is a Distributed Denial of Service (DDoS) attack detection and mitigation device that cleans the zone traffic as the traffic flows through it, dropping the attack traffic and injecting the legitimate traffic back into the network. When the Detector module determines that the zone is under attack, it can activate the Guard attack mitigation services. The Detector module can also synchronize zone configurations with the Guard. For more information about the Guard, see the Cisco Anomaly Guard Module Configuration Guide or the Cisco Guard Configuration Guide.

This chapter contains the following sections:

Understanding Zone Policies

Understanding Policy Templates

Understanding the Policy Path

Configuring Policy Parameters

Understanding Worm Policies

Monitoring Policies

Backing Up the Policy Configuration

Understanding Zone Policies

The zone policies enable the Detector module to perform a statistical analysis of the zone traffic flow. The zone policies are configured to take action against a particular traffic flow if the flow exceeds the policy thresholds, indicating malicious or abnormal traffic. When a flow exceeds the policy thresholds, the policies configure a set of filters (dynamic filters) dynamically to record the event in its syslog or activate a Guard that you have defined in the remote Guard lists. When activated, the Guard protects the zone by mitigating the attack.

Every zone configuration contains a set of policies. When you create a new zone using a predefined zone template, the Detector module configures the new zone with policies associated with the template. When you create a new zone by copying an existing zone, the Detector module configures the new zone with the policies of the existing zone.

To create zone-specific policies and tune their thresholds to recognize normal zone traffic, the Detector module learns the zone traffic in a two-phase learning process (see "Understanding the Learning Process" section on page 1-4). The Detector module uses predefined policy templates to construct the policies and then learns the policy thresholds as determined by the zone traffic. The Detector module uses each policy template to create policies that the Detector module requires to protect the zone against a specific Distributed Denial of Service (DDoS) threat. After the Detector module creates and tunes the zone policies, you can add and delete policies or change policy parameters.

Policies have cross dependencies and priorities. If two different policies define the same traffic flow, the Detector module analyzes the flow using the policy that is more specific. For example, policies relating to TCP services exclude the HTTP services that are handled by the HTTP-related policies.

You can configure the policy triggers and the action that the policy takes once it is activated.

Understanding Policy Templates

A policy template is a collection of policy construction rules that the Detector module uses during the policy construction phase to create the zone policies. The name of the policy template is derived from the characteristics that are common to all the policies that it creates and can be a protocol (such as DNS), an application (such as HTTP), or the objective (such as ip_scan). For example, the policy template tcp_connections produces policies that relate to a connection, such as the number of concurrent connections. When you create a new zone, the Detector module includes a set of policy templates in the zone configuration.

This section contains the following topics:

Understanding the Different Policy Template Types

Configuring Policy Template Parameters

Understanding the Different Policy Template Types

Table 7-1 describes the Detector module policy templates. The Detector module includes these policy templates when you create a new zone using the DETECTOR_DEFAULT zone template.

Table 7-1 Policy Templates 

Policy Template
Constructs a Group of Policies Relating To

dns_tcp

DNS-TCP protocol traffic.

dns_udp

DNS-UDP protocol traffic.

fragments

Fragmented traffic.

http

HTTP traffic that flows, by default, through port 80 (or other user-configured ports).

ip_scan

IP scanning. A situation in which a client from a specific source IP address tries to access many destination IP addresses in the zone. This policy template is designed primarily for zones in which the IP address definition is a subnet.

By default, this policy template is disabled. The default action for this policy template is notify.

Note The policies that are produced from this policy template consume system resources and can affect the performance of the Detector module.

other_protocols

Non-TCP and non-UDP protocols.

port_scan

Port scanning. A situation in which a client from a specific source IP address tries to access many ports in the zone.

By default, this policy template is disabled. The default action for this policy template is notify.

Note The policies that are produced from this policy template consume system resources and can affect the performance of the Detector module.

tcp_connections

TCP connection characteristics.

tcp_not_auth

TCP connections that have not been authenticated by the Detector module anti-spoofing functions.

tcp_outgoing

TCP connections initiated by the zone.

tcp_ratio

Ratios between different types of TCP packets, for example, the ratio of SYN packets to FIN/RST packets.

tcp_services

TCP services on ports other than HTTP-related, such as ports 80 and 8080.

udp_services

UDP services.


The Detector module includes additional policy templates for zones that were created from zone templates that are designed for specific types of attacks or specific services. Table 7-2 details the policy templates that the Detector module adds to a zone configuration based on a specific zone template.

Table 7-2 Additional Policy Templates 

Zone Template
Policy Template

DETECTOR_WORM

worm_tcp—Constructs a group of policies that identify TCP worms. Worm TCP policies manage worm attacks in which one or more source IP addresses create many nonestablished connections on the same port to many destination IP addresses. This policy template is designed primarily for zones in which the IP address definition is a subnet.

The Detector module adds services to policies that are created from this policy template during the threshold tuning phase of the learning process instead of during the policy construction phase. The policy template parameters, max_services and min_threshold, do not apply to this policy template. See the "Understanding Worm Policies" section for more information.


If you create a zone from a GUARD_ zone template, you can configure the parameters of additional policy templates that can be synchronized to a Guard. The Detector module uses the policy templates described in Table 7-3 and replaces the policy templates http, tcp_connections, and tcp_outgoing with the policy templates http_ns, tcp_connections_ns, and tcp_outgoing_ns policies. The http_ns, tcp_connections_ns, and tcp_outgoing_ns policy templates do not create policies with actions that require the Guard to apply the strong protection level to the traffic flow.

Table 7-3 details the Detector module policy templates for GUARD_TCP_NO_PROXY.

Table 7-3 GUARD_TCP_NO_PROXY Policy Templates 

Policy Template
Replaces Policy Template
Constructs a group of policies relating to

tcp_connections_ns

tcp_connections

TCP connection characteristics.

tcp_outgoing_ns

tcp_outgoing

TCP connections initiated by the zone.

http_ns

http

HTTP traffic flowing, by default, through port 80 (or other user-configured ports).


To view a list of all policy templates, use the policy-template command in zone configuration mode and press Tab twice.

Configuring Policy Template Parameters

During the learning process, each active policy template produces a group of policies based on the policy definitions and the zone traffic characteristics. The Detector module ranks the services (protocol and port numbers) that the policy template monitors by the level of traffic volume. The Detector module then selects the services that have the highest traffic volume and that have exceeded the defined minimum threshold, and it creates a policy for each service. Some of the policy templates create an additional policy to handle all traffic flows for which a specific policy was not added with a service of any.

You can configure the following policy template parameters:

Maximum Number of Services—Defines the maximum number of services that the Detector module picks up for the policy template to create specific policies.

Minimum Threshold—Defines the minimum threshold that must be exceeded for the Detector module to rank the service.

Policy Template State—Defines whether or not the Detector module produces policies from the policy template.

The policy template parameters maximum number of services and minimum threshold do not affect the worm_tcp policy template.

To configure the policy template parameters, enter the policy template configuration mode by entering the following command in zone configuration mode:

policy-template policy-template-name

The policy-template-name argument specifies the name of the policy template. See Table 7-1 for more information.

After executing the command, the Detector module enters the policy template configuration mode.

The following example shows how to enter http policy template configuration mode:

user@DETECTOR-conf-zone-scannet# policy-template http
user@DETECTOR-conf-zone-scannet-policy_template-http#

To display the parameters of a specific policy template, use the show command in policy template configuration mode.

This section contains the following topics:

Configuring the Maximum Number of Services

Configuring the Minimum Threshold

Configuring Policy Template States

Configuring All Policy Template Parameters Simultaneously

Configuring the Maximum Number of Services

The maximum number of services parameter defines the maximum number of services (protocol numbers or port numbers) for which the policy template selects and creates policies. The Detector module ranks the services by the level of traffic volume for each service. The Detector module then selects the services that have the highest traffic volume and that have exceeded the defined minimum threshold (as defined by the min-threshold parameter), and it creates policies for each service. The Detector module may add an additional policy with a service of any to handle all other traffic flows with the characteristics of the policy template.


Note The higher the maximum number of services, the more Detector module memory the zone requires.


You can only define the maximum number of services parameter for policy templates that detect services: tcp_services, tcp_services_ns, udp_services, and other protocols. You cannot configure it for policy templates that monitor a specific service, such as dns_tcp, which monitors service 53, or for policy templates that relate to a specific traffic characteristic, such as fragments.

The Detector module measures the traffic rate of the service based on the policy traffic characteristics. The traffic characteristic can be the source IP addresses or the destination IP addresses. A policy that monitors the service any measures the rate of source IP addresses on all services that are not handled by a specific policy.

By limiting the service number, you can configure the Detector module policies to your preferred traffic flow requirements.

To configure the maximum number of services, use the following command in policy template configuration mode:

max-services max-services

The max-services argument is an integer greater than 1 that defines the maximum number of services that the Detector module selects. We recommend that you do not exceed the maximum of 10 services.

The following example shows how to configure the maximum number of services that the Detector module monitors to 5:

user@DETECTOR-conf-zone-scannet-policy_template-tcp_services# max-services 5

Configuring the Minimum Threshold

The minimum threshold parameter defines the minimum traffic volume for a service. When the threshold is exceeded, the Detector module constructs policies that relate to the service traffic according to the particular traffic flow that exceeded the threshold. By setting the threshold, you can adapt the anomaly detection operation to the traffic volume of the zone services.

You cannot configure the minimum threshold parameter for policy templates that are essential for proper zone anomaly detection and always construct a policy such as the following policy templates: tcp_services, udp_services, other_protocols, http, and fragments.

To configure the minimum threshold, use the following command in policy template configuration mode:

min-threshold min-threshold

The threshold argument is a real number (a floating point number with two decimal places), equal to or greater than 0, that defines the minimum threshold rate in packets per second (pps). When measuring concurrent connections and the SYN/FIN ratio, the threshold is an integer that defines the total number of connections.

The following example shows how to configure the minimum threshold of the policy template http:

user@DETECTOR-conf-zone-scannet-policy_template-http# min-threshold 12.3

Configuring Policy Template States

The policy template state parameter defines whether the policy template is enabled or disabled. The Detector module is unable to use a disabled policy template during the policy construction phase to produce policies.


Caution Disabling a policy template may seriously compromise zone anomaly detection. If you disable a policy template, the Detector module cannot detect the zone traffic to which the policy template relates. For example, disabling the dns_udp policy template prevents the Detector module from creating zone policies that manage DNS (UDP) attacks.

To disable a policy template, use the disable command in policy template configuration mode.

To enable a policy template, use the enable command in policy template configuration mode.

The following example shows how to disable the policy template http:

user@DETECTOR-conf-zone-scannet-policy_template-http# disable

Configuring All Policy Template Parameters Simultaneously

You can configure all policy template operational parameters with a single command by entering the following command in zone configuration mode:

policy-template policy-template-name max-services min-threshold {disabled | enabled}

Table 7-4 provides the arguments and keywords for the policy-template command.

Table 7-4 Arguments and Keywords for the policy-template Command 

Parameter
Description

policy-template-name

Policy template name. See Table 7-5 for more information.

max-services

Maximum number of services for which the Detector module selects and constructs policies from the specific policy template.

To prevent the Detector module from changing the current value, enter a value of -1.

See the "Configuring the Maximum Number of Services" section for more information.

min-threshold

Minimum threshold that must be exceeded for the Detector module to rank the service.

To prevent the Detector module from changing the current value, enter a value of -1.

See the "Configuring the Minimum Threshold" section for more information.

disabled

Disables the policy template from producing policies. See the "Configuring Policy Template States" section for more information.

enabled

Enables the policy template. See the "Configuring Policy Template States" section for more information.


The following example shows how to set the parameters of the tcp_services policy template. The maximum number of services is set to 3, the policy state is set to enabled, and the minimum threshold is unchanged (-1).

user@DETECTOR-conf-zone-scannet# policy-template tcp_services 3 -1 enabled

Understanding the Policy Path

The name of the policy is composed of sections that describe the traffic characteristic that it measures. For example, the policy http/80/analysis/syns/src_ip measures traffic flows of HTTP SYN packets destined to port 80 that were authenticated by the Detector module analysis detection level functions and aggregated according to source IP addresses.

Figure 7-1 provides an example of a zone policy name.

Figure 7-1 Policy Name

Table 7-5 describes the policy name sections.

Table 7-5 Policy Name Sections 

Section
Description

Policy template

Policy template that was used to construct the policy. Each policy template deals with the characteristics that the Detector module requires to detect a specific DDoS threat. See the "Understanding Policy Templates" section for more information.

Service

Port number or protocol number in the traffic flow that the policy monitors.

Detection level

Detection level that the Detector module applies to the traffic flow. Detection levels have a static configuration and cannot be configured manually.

Packet types

Packet types that the Detector module monitors.

Traffic characteristics

Traffic characteristics that the Detector module uses to aggregate the policy.


The first four sections of the policy name (policy template, service, detection level, and packet type) define the type of traffic that is analyzed. The last section of the policy path (traffic characteristics) defines how to analyze the flow.

This section describes each of the policy path sections as follows:

Understanding and Managing the Policy Services

Understanding the Packet Types that the Detector module Monitors

Understanding the Traffic Characteristics that the Detector Module Monitors

Understanding and Managing the Policy Services

The service section defines the zone application port or protocol to which each policy relates. Policies have cross dependencies and priorities. If two different policies define the same traffic flow, the Detector module analyzes the flow using the policy that is more specific. The service any relates to all traffic that does not specifically match other services created from the same policy template.

We recommend that you define specific policies for the zone main services to obtain anomaly detection that is most suited to your individual needs.


Caution Do not add the same service (port number) to more than one policy because it may decrease the Detector module performance.

When you add or delete a service from the zone policies, the Detector module marks the zone policies as untuned. If you enabled zone anomaly detection and the learning process, the Detector module cannot detect anomalies in the zone traffic until you perform one of the following actions:

Perform the threshold tuning phase of the learning process and accept the results (see the "Activating the Threshold Tuning Phase" section on page 8-6).

Mark the zone policies tuned (see the "Marking the Policies as Tuned" section on page 8-10).

This section contains the following topics:

Adding a Service

Deleting a Service

Adding a Service

You can add services to all policies that were created from a specific policy template. The new service is an addition to the services that were discovered during the policy construction phase and is defined with default values. You can define the threshold manually, but we recommend that you run the threshold tuning phase of the learning process to tune the policies to the zone traffic. See the "Activating the Threshold Tuning Phase" section on page 8-6 for more information.

You can add a new service to policies that were created from the following policy templates:

tcp_services, udp_services, tcp_services_ns, or worm_tcp

The service designates a port number.

other_protocols

The service designates a protocol number.


Note If you activate the policy construction phase after adding a service, new services might override the manually added service.


Unless you enable the policy construction phase, you may need to add a service manually in the following situations:

A new application or service was added to the zone network.

The policy construction phase was activated for a short period, so it does not reflect all the network services (for instance, if there are known applications or services that are active only once a week or during the night).

To add a service, use one of the following commands:

add-service service-num (in policy template configuration mode)

policy-template policy-template-name add-service service-num (zone configuration mode)

Table 7-6 provides the arguments for the add-service command.

Table 7-6 Arguments for the add-service Command 

Parameter
Description

service-num

Protocol or port number.

policy-template-name

Policy template name. See Table 7-1 for more information.


The following example shows how to add a service to all the policies that were created from the policy template tcp_services:

user@DETECTOR-conf-zone-scannet-policy_template-tcp_services# add-service 25

Deleting a Service

You can delete a specific service for any policy template. The Detector module will delete the service from all policies that were created from the specific policy template.

To delete a service, use one of the following commands:

remove-service service-num (in policy template configuration mode)

policy-template policy-template-name remove-service service-num (in zone configuration mode)

Table 7-7 provides the arguments for the remove-service command.

Table 7-7 Arguments for the remove-service Command 

Parameter
Description

service-num

Protocol or port number to remove.

policy-template-name

Policy template name. See Table 7-1 for more information.



Caution If you delete a service, the Detector module policies cannot monitor the traffic of that service, which may compromise zone anomaly detection.

You can remove services from the following policy templates:

tcp_services, udp_services, or tcp_services_ns

The service is a port number.

other_protocols

The service is a protocol number.

If you do not activate the policy construction phase of the learning process, you may need to remove a service manually in the following situations:

An application or service was removed from the network.

An application or service that you do not want to enable (because it is uncommon for the network environment) but was identified during the policy construction phase.


Note If you activate the policy construction phase after removing a service, the same service might be added again.


The following example shows how to delete a service from all policies that were created from the policy template tcp_services:

user@DETECTOR-conf-zone-scannet-policy_template-tcp_services# remove-service 25

Understanding the Packet Types that the Detector module Monitors

The Detector module monitors packet characteristics, which can be one of the following:

Packet type (for example, TCP-SYN packets)

Packet analysis (for example, authenticated packets, which are packets that the Detector module has verified their connection by performing a TCP handshake)

Packet direction (for example, incoming connections)

Table 7-8 describes the packet types that the Detector module monitors.

Table 7-8 Packet Types 

Packet Type
Description

auth_pkts

Packets for which either a TCP handshake or UDP authentication was performed.

auth_tcp_pkts

Packets for which a TCP handshake was performed.

auth_udp_pkts

Packets for which UDP authentication was performed.

in_nodata_conns

Incoming zone connections that have no data transfer on the connection (packets without a data payload).

in_conns

Incoming zone connections.

in_pkts

Incoming zone DNS query packets.

in_unauth_pkts

Incoming zone unauthenticated DNS queries.

non_estb_conns

Nonestablished connections. Incoming zone failed connections, which are TCP connection requests (SYN packets) for which no reply was received.

out_pkts

Zone incoming DNS reply packets.

reqs

Request packets with a data payload.

syns

Synchronization packets (TCP SYN flagged packets).

syn_by_fin

SYN and FIN flagged packets. The Detector module verifies the ratio between the number of SYN flagged packets and the number of FIN flagged packets.

unauth_pkts

Packets that did not undergo a TCP handshake.

pkts

All packet types that do not fall under any other category in the same detection level.


Understanding the Traffic Characteristics that the Detector Module Monitors

Traffic characteristics define how to analyze the traffic flow and what characteristics were used to aggregate the policies. Different policies can analyze the same traffic flow but measure the rate based on different characteristics, as shown in this example:

dns_tcp/53/analysis/pkts/dst_ip and dns_tcp/53/analysis/pkts/src_ip.

Table 7-9 describes the traffic characteristics that the Detector module monitors.

Table 7-9 Traffic Characteristics 

Traffic Characteristic
Description

dst_ip

Traffic destined to a zone IP address.

dst_ip_ratio

Ratio of SYN and FIN flagged packets destined to a specific IP address.

dst_port

Traffic destined to a specific zone port.

dst_port_ratio

Ratio of SYN and FIN flagged packets destined to a specific port.

global

Summation of all traffic flow as defined by the other policy sections.

protocol

Traffic destined to the zone aggregated based on the protocol.

scanners

Histogram of the number of source IP addresses that scan zone destination IP addresses on a specific destination port. See the "Understanding Worm Policies" section for more information.

src_ip

Traffic destined to the zone aggregated according to the source IP address.

src_ip_many_dst_ips

Traffic from a single IP address that probes a large number of zone IP addresses on the same port. This key is used for IP scanning.

src_ip_many_ports

Traffic from a single IP address that probes a large number of ports on a zone destination IP address. This key is used for port scanning.


Configuring Policy Parameters

After completing the learning process, you can display specific policy parameters (policy state, policy threshold, policy timeout, policy action, and policy interactive state) to determine if the policy parameters suit the zone traffic. You can configure the policy parameters of a single policy or a group of policies to adapt to zone traffic requirements.

To display the configuration of the policy parameters, use the show command in policy configuration mode.

To enter policy configuration mode, use the following command in zone configuration mode:

policy policy-path

The policy-path argument specifies the policy path sections. The path can be a partial path that includes only part of the policy sections. See the "Understanding Zone Policies" section for more information.


Note To move up one level in the policy path hierarchy, enter policy .. at the policy path prompt.


The following example shows how to enter the dns_tcp/53/analysis/syns/global policy configuration mode:

user@DETECTOR-conf-zone-scannet# policy dns_tcp/53/analysis/syns/global 
user@DETECTOR-conf-zone-scannet-policy-/dns_tcp/53/analysis/syns/global#

You can change the policy action, timeout, threshold, and learning parameters at every section of the policy path. However, more policies are affected if you change these parameters at the higher-level policy sections (such as policy template or service sections). If you configure these parameters at a high-level policy path hierarchy, these parameters change in all the subpolicy paths.

You can use an asterisk (*) as a wildcard character in each policy path section. If you do not specify a policy path section, the Detector module relates to the unspecified section as a wildcard (*). For example, the policy tcp_services//analysis//global uses a wildcard for the service and the packet type.

This section contains the following topics:

Changing the Policy State

Configuring the Policy Threshold

Configuring the Policy Timeout

Configuring the Policy Action

Configuring the Policy Interactive Status

Changing the Policy State

The zone policies have three possible states:

Active—The policy monitors the traffic and performs an action once the threshold is exceeded.

Inactive—The policy monitors the traffic and obtains the threshold, but it takes no action when a threshold is exceeded. You can inactivate a policy to avoid reactivating the threshold-tuning phase of the learning process.

Disabled—The policy does not monitor the traffic flow, so no threshold is obtained.


Note We recommend that you activate the threshold tuning phase of the learning process to ensure that the Detector module monitors the correct thresholds for the other policies.



Caution When you disable a policy, the active zone policies assume responsibility for the traffic that would normally be monitored by the disabled policy. To adjust the thresholds of the active policies, we recommend that you activate the threshold tuning phase before you activate zone anomaly detection.

To change the policy state, use the following command in policy configuration mode:

state {active | disabled | inactive}

The following example shows how to set the policy state:

user@DETECTOR-conf-zone-scannet-policy-/dns_tcp/53/analysis/syns# state disabled

The following example shows how to set the state of all global policies:

user@DETECTOR-conf-zone-scannet-policy-/*/*/*/global# state inactive


Caution If you deactivate or disable a policy, the zone policies may not assume their role, and the zone anomaly detection can be compromised.

If you activate the policy construction phase after disabling a zone policy, all zone policies are reconfigured according to the current traffic flow and the policy may be reactivated.

Configuring the Policy Threshold

The policy threshold defines the threshold traffic rate for a specific policy and is adjusted by the threshold tuning phase. When this threshold is exceeded, the policy performs the action that is defined by the policy action.

The threshold is measured in packets per second except for policies that are constructed from the following policy templates:

num_soruces—The threshold is measured in the number of IP addresses or ports.

tcp_connections—The threshold is measured in the number of connections.

tcp_ratio—The threshold is measured as the ratio number.

worm_tcp—The threshold is measured as the maximum number of zone destination IP addresses that a source IP may scan.

You can configure the policy threshold in the following ways:

Set the threshold—You can set the value of the policy threshold. See the "Setting the Policy Threshold" section.

Multiply the threshold—The Detector module multiplies the current policy thresholds by a factor. The new value may change in subsequent threshold tuning phases if you do not set it as fixed. See the "Multiplying a Threshold by a Factor" section.

Configure specific IP thresholds—The Detector module sets thresholds for specific IP source addresses within the zone address range. See the "Configuring Specific IP Thresholds" section.

The policy threshold may change if you perform additional threshold tuning phases. You can modify how a threshold may change in subsequent threshold tuning phases in the following ways:

Set the threshold as fixed—The Detector module will not change the value of the policy threshold, proxy-threshold, and threshold-list in subsequent threshold tuning phases. See the "Setting the Threshold as Fixed" section.

Set a fixed multiplier for the policy threshold—The Detector module calculates the policy threshold in subsequent threshold tuning phases based on the current policy threshold, the learned threshold, and the fixed multiplier. See the "Configuring a Threshold Multiplier" section.

This section contains the following topics:

Setting the Policy Threshold

Setting the Threshold as Fixed

Configuring a Threshold Multiplier

Multiplying a Threshold by a Factor

Configuring Specific IP Thresholds

Setting the Policy Threshold

To configure the policy threshold, use the following command in policy configuration mode:

threshold threshold

The threshold argument is a positive number that specifies the policy threshold.

The following example shows how to set the threshold value of the policy dns_tcp/53/analysis/syns/global to 300:

user@DETECTOR-conf-zone-scannet-policy-/dns_tcp/53/analysis/syns/ 
global# threshold 300

Setting the Threshold as Fixed

You can set a policy threshold, proxy-threshold, and threshold-list as fixed. The Detector module ignores new thresholds in the threshold tuning phase of the learning process and maintains the current thresholds. Setting a threshold as fixed enables you to configure the thresholds of a policy but continue learning the thresholds of other policies.

To set a policy threshold as fixed, use the following command in policy configuration mode:

learning-params fixed-threshold

The following example shows how to set the threshold of the policy dns_tcp/53/analysis/syns/global as fixed:

user@DETECTOR-conf-zone-scannet-policy-/dns_tcp/53/analysis/syns/ 
global# learning-params fixed-threshold

You can set the threshold of several policies as fixed in a single command by entering the command in zone configuration mode. To set a policy threshold as fixed while in zone configuration mode, use the following command:

policy policy-path learning-params fixed-threshold

The policy-path argument specifies the policy path. The path can be a partial path that includes only part of the policy sections. See the "Understanding Zone Policies" section for more information.

The following example shows how to set the thresholds of all policies that were created from the dns_tcp policy template as fixed:

user@DETECTOR-conf-zone-scannet# policy dns_tcp learning-params fixed-threshold

To display the policy learning parameters, use the show learning-params command in policy configuration mode, or use the show policies policy-path learning-params command in zone configuration mode.

Configuring a Threshold Multiplier

You can set a multiplier for a policy threshold. The Detector module calculates a new policy threshold by multiplying the learned threshold by the specified multiplier before accepting the results of subsequent threshold tuning phases. The Detector module accepts the results of the threshold tuning phase using the configured threshold selection method. See the "Configuring the Threshold Selection Method" section on page 8-9.

To set a multiplier for the policy threshold, use the following command in zone configuration mode:

policy policy-path learning-params threshold-multiplier threshold-multiplier

Table 7-10 provides the arguments and keywords for the policy learning-params threshold-multiplier command.

Table 7-10 Arguments and Keywords for the policy learning-params threshold-multiplier Command 

Parameter
Description

policy-path

Policy path for which to multiply the thresholds. The path can be a partial path that includes only part of the policy sections. See the "Understanding Zone Policies" section for more information.

learning-params

Configures the learning parameters.

threshold-multiplier threshold-multiplier

Multiplies the policy threshold. The threshold-multiplier is a real positive number (a floating point number with two decimal places) by which the policy threshold is multiplied. Enter a number less than 1 to decrease the policy threshold.


To set a multiplier for the policy threshold in policy configuration mode, use the learning-params threshold-multiplier threshold-multiplier command.

The following example shows how to configure a threshold multiplier so that the Detector module decreases the thresholds of policies that were created from the policy template dns_tcp by half in subsequent threshold tuning phases:

user@DETECTOR-conf-zone-scannet# policy dns_tcp learning-params threshold-multiplier 0.5

To display the policy learning parameters, use the show learning-params command in policy configuration mode, or use the show policies policy-path learning-params command in zone configuration mode.

Multiplying a Threshold by a Factor

You can multiply the thresholds of a policy or a group of policies by a factor, which enables you to increase or decrease the threshold of a policy or a group of policies if the traffic volume does not represent the zone traffic. You can enable the Detector module to multiply the policy thresholds, the proxy thresholds, and the thresholds that were defined by the policy threshold-list command.

To multiply policy thresholds by a factor, use the following command in zone configuration mode:

policy policy-path thresh-mult threshold-multiply-factor

Table 7-11 provides the arguments and keywords for the policy thresh-mult command.

Table 7-11 Arguments and Keywords for the policy thresh-mult Command 

Parameter
Description
policy-path

Policy template name. See Table 7-1 for more information.

thresh-mult threshold-multiply-factor

Specifies a real positive number (a floating point number with 4 decimal places) by which to multiply the threshold. Enter a number less than 1 to decrease the policy threshold.


The following example shows how to decrease the thresholds of policies that were created from the policy template dns_tcp by half:

user@DETECTOR-conf-zone-scannet# policy */*/*/*/src_ip thresh-mult 0.5


Note The Detector module may change the threshold value in subsequent threshold tuning phases. To prevent the Detector module from changing the threshold value, set the threshold value as fixed. See the "Setting the Threshold as Fixed" section.


To display the policy learning parameters, use the show learning-params command in policy configuration mode, or use the show policies policy-path learning-params command in zone configuration mode.

Configuring Specific IP Thresholds

You can avoid false attack detections by the Detector module when traffic increases on a known high traffic source or destination IP address by configuring a policy with a threshold for traffic that is associated with that IP address.

You should consider configuring a specific IP threshold if one of the following situations occurs:

When there is known high-volume traffic from a source IP address, you can configure a threshold to apply to traffic that originates from the specific source IP address.

When there is a nonhomogeneous zone (a zone that consists of more than a single IP address) and there is known high-volume traffic flowing to part of the zone only, you can configure a threshold to apply to traffic that targets the specific destination IP address within the zone.

You can configure a specific IP threshold only for policies with traffic characteristics of destination IP (dest_ip).

To configure a specific IP threshold, use one of the following commands:

policy policy-path threshold-list ip threshold [ip threshold ...] (in zone configuration mode)

threshold-list ip threshold [ip threshold ...] (in policy configuration mode)

Table 7-12 provides the arguments for the threshold-list command.

Table 7-12 Arguments for the policy threshold-list Command 

Parameter
Description
policy-path

Policy template name. See Table 7-1 for more information.

ip

Specific IP address.

threshold

Threshold traffic rate in packets per second, except for policies that measure concurrent connections and SYN-by-FIN ratio where the threshold is the number of connections.


You can add a maximum of 10 specific IP thresholds for each policy. You can enter all specific IP thresholds in a single command.

The Detector module might change the policy thresholds in subsequent threshold tuning phases if the threshold selection method is set to new-thresholds. See the "Configuring the Threshold Selection Method" section on page 8-9 for more information.

The following example shows how to set specific IP thresholds for IP addresses 10.10.10.2 and 10.10.15.2 for the policy http/80/analysis/syns/src_ip:

user@DETECTOR-conf-zone-scannet-policy-/http/80/analysis/syns/src_ip# threshold-list 
10.10.10.2 500 10.10.15.2 500

Configuring the Policy Timeout

The timeout parameter defines the minimum time for dynamic filters that are produced by the policy to apply their action.

To configure the policy timeout, use the following command in policy configuration mode:

timeout {forever | timeout}

Table 7-13 provides the arguments and keywords for the timeout command.

Table 7-13 Arguments and Keywords for the timeout Command 

Parameter
Description
forever

Specifies an indefinite time span.

timeout

Integer from 1 to 3,000,000 that specifies the minimum time in seconds that the dynamic filters, which are produced by the policy, are active.


The following example shows how to set the timeout of the policy http/80/analysis/syns/src_ip to 100 seconds:

user@DETECTOR-conf-zone-scannet-policy-/http/80/analysis/syns/src_ip# timeout 100

To change the timeout of a group of policies simultaneously, use the policy set-timeout command in zone configuration mode.

The following example shows how to set the timeout of all policies that were produced from the HTTP policy template and measure source IP addresses to 100:

user@DETECTOR-conf-zone-scannet# policy http/*/*/*/src_ip set-timeout 100

Configuring the Policy Action

The action parameter defines the type of action that the policy takes once its threshold is exceeded.

To configure the policy action, use the following command in policy configuration mode:

action policy-action

Table 7-14 describes the policy actions.

Table 7-14 Policy Actions 

Policy Action
Description

notify

Notifies you when its threshold is exceeded.

remote-activate

Activates remote Guards when its threshold is exceeded. The remote Guards are defined in the remote Guard lists. See the "Activating Remote Guards to Protect a Zone" section on page 9-5 for more information.


The following example shows how to set the action of the policy http/80/analysis/syns/src_ip:

user@DETECTOR-conf-zone-scannet-policy-/http/80/analysis/syns/src_ip# action 
remote-activate

To change the action of a group of policies simultaneously, use the policy set-action command in zone configuration mode.

The following example shows how to set the action of all dns_tcp policies:

user@DETECTOR-conf-zone-scannet# policy dns_tcp/ set-action remote-activate 
set action of dns_tcp/ to remote-activate: 
4 policy actions set.

Configuring the Policy Interactive Status

The interactive status parameter defines the interactive status that the pending dynamic filters, which are created by the zone policy, will assume. The interactive status applies only to zones if you enable zone anomaly detection, and the zone is in interactive detect mode. See Chapter 10, "Using Interactive Detect Mode," for more information.

To modify the status of the pending dynamic filters that a policy produces after you have set the interactive status of a recommendation to always-accept or always-ignore, use the interactive-status command.

For example, if you have defined the status of a recommendation to always-accept, the recommendation and the pending dynamic filters of the recommendation are no longer displayed. To ignore the recommendation or the pending dynamic filters that the recommendation produces, change the policy interactive status to interactive or always-accept.

To configure the policy interactive status, use the following command in policy configuration mode:

interactive-status {always-accept | always-ignore | interactive}

Table 7-15 provides the keywords for the interactive-status command.

Table 7-15 Keywords for the interactive-status Command 

Parameter
Description

always-accept

Accepts the dynamic filters that the policy produces automatically. The action applies automatically whenever the policy produces new recommendations.

The Detector module does not display these recommendations.

always-ignore

Ignores the dynamic filters that the policy produces automatically. The policy does not produce recommendations when its threshold is exceeded.

The Detector module does not display these recommendations.

interactive

Waits for you to accept or ignore the dynamic filters that the policy produces.

The Detector module displays these dynamic filters as part of the recommendations.


The following example shows how to configure the interactive status of policy dns_tcp/53/analysis/pkts/src_ip to always-accept:

user@DETECTOR-conf-zone-scannet-policy-/dns_tcp/53/analysis/pkts/ 
src_ip# interactive-status always-accept

Understanding Worm Policies

Internet worms are automated, self-propagating, intrusive agents that make copies of themselves and facilitate their distribution. Worms attack a vulnerable host, infect it, and then use it as a base to attack other vulnerable targets. They search for other targets by using a form of network inspection, typically a scan, and propagate to the next target. A scanning worm locates vulnerable hosts by generating a list of addresses to probe and then contacting the hosts. Code Red worm, Sasser worm, Blaster worm, and the Slammer worm are all examples of high-profile worms that spread in this manner.

The Detector module enables you to detect TCP worm attacks by identifying worms through abnormal traffic patterns that indicate that the zone network is being scanned. The Detector module assumes that even if no TCP worm attack is in progress, there may be some scanners in the network. It identifies a scanner as a source IP address that is the initiator of nonestablished connections (an incoming SYN packet for which no SYN/ACK reply packet was identified) to many zone destination IP addresses on a specific port.

To analyze the zone traffic, the Detector module uses a table that holds frequency data, which is known as a histogram, of network scanners. The Detector module first learns the zone network when no attack is in progress, and then it creates a histogram of concurrent scanners. The histogram describes the number of scanners that concurrently scan specific numbers of zone destination IP addresses. The Detector module then measures how many scanners access more than a specific number of zone destination IP addresses.

The Detector module uses two types of thresholds to analyze worm traffic characteristics:

Scanning threshold—Defines the maximum number of zone IP addresses that a single source IP address may scan. This threshold is defined by the policy threshold.

Histogram threshold—Defines the maximum number of source IP addresses that can scan more than the specified numbers of zone IP addresses.

The Detector module identifies a worm attack when there is a deviation from the histogram that it has learned when no attack was in progress (that is, when the number of source IP addresses that concurrently scan more than the defined zone destination IP addresses is exceeded). See the "Identifying Worm Attacks" section for more information.

Worm policies differ from other policies as follows:

The Detector module learns new services for worm policies during the threshold tuning phase, rather than during the policy construction phase, so you may see new services (ports) added to worm policies during the threshold tuning phase.

The service any relates to ports for which the Detector module does not have specific policies. For example, if the Detector module has policies for worm_tcp/80 and worm_tcp/50, the policy worm_tcp/any monitors all traffic that is not destined to ports 50 or 80. Unlike other policies, the any service does not aggregate the traffic to all unspecified ports. When the Detector module monitors the zone traffic, it holds a separate, internal histogram for each port that is scanned. It compares this histogram with the histogram of the any service.

This section contains the following topics:

Configuring Worm Policies

Identifying Worm Attacks

Configuring Worm Policies

The worm_tcp policy template is available in the DETECTOR_WORM zone template only.

The policies that manage TCP worms are constructed from the worm_tcp policy template, the non_estb_conns packet type, and the scanner's traffic characteristics.

You can configure the histogram and change the scanning thresholds by entering the following command in policy configuration mode:

histogram num-dst-ips num-src-ips [num-dst-ips num-src-ips...]

Table 7-16 provides the arguments for the histogram command.

Table 7-16 Arguments for the histogram Command 

Parameter
Description

num-dst-ips

Number of scanned zone destination IP addresses. The values of num-dst-ips are 5, 20, and 100 and are system defined. You can modify the value of the num-src-ips that is defined for each num-dst-ips.

num-src-ips

Histogram threshold. When the threshold is exceeded, the policy takes the action that is defined by the policy action parameter. The threshold specifies the number of source IP addresses that may scan the specified number of zone destination IP addresses (num-dst-ips).


You can enter all the histogram thresholds in a single command.

The following example shows how to set the histogram thresholds for all frequencies:

user@DETECTOR-conf-zone-scannet- worm_tcp/445/analysis/non_estb_conns/scanners# histogram 
5 99 20 80 50 8 100 1

To display the current histogram settings, use the show policies command.

You can set the maximum number of zone IP addresses that a single source IP address may scan (scanning threshold). To set this number, use the threshold command. See the "Configuring the Policy Threshold" section for more information.

To specify the histogram thresholds for a specific port, use the add-service command to add a service for the specific port number to all policies that were created from the worm_tcp policy template. See the "Adding a Service" section for more information.

Identifying Worm Attacks

The Detector module uses two types of thresholds to analyze worm traffic characteristics: a scanning threshold and a histogram threshold. See the "Understanding Worm Policies" section for more information.

When a histogram threshold is exceeded, the Detector module produces a dynamic filter with an unspecified source IP address (*). This dynamic filter indicates that a worm attack is in progress. The dynamic filter policy threshold specifies which histogram threshold was exceeded. The Detector module defines a new, internal scanning threshold that is equal to the dynamic filter policy threshold.

The source IP addresses that scan the zone destination IP addresses are those of worm-infected hosts. As long as the zone is under attack, each worm-infected host that scans more zone destination IP addresses than the maximum defined by the new, internal scanning threshold causes the production of a dynamic filter. The Detector module acts on these attacking flows as defined by the dynamic filter action.

For example, if the policy threshold (the scanning threshold) is 300, and the policy scanner's histogram for port 445 is as shown in Table 7-17, then if the Detector module identifies a scanner that scans 350 zone destination IP addresses, it produces a dynamic filter indicating that a mass scanner was detected. However, this scanner does not yet imply that a worm attack is in progress.

Table 7-17 Sample Histogram

Number of source IP addresses
10
5
2
Number of Destination IP addresses
5
20
100

When the Detector module identifies six concurrent source IP addresses that scan more than 50 zone destination IP addresses on port 445, it produces a dynamic filter from the worm_tcp policy with an unspecified source IP address (*) that indicates that the Detector module has identified a worm attack on port 445. The dynamic filter policy threshold, 50, specifies the new internal scanning threshold and causes the Detector module to lower the threshold definition of a scanner, so that the Detector module produces additional dynamic filters for each source IP address that scans more than the new scanning threshold (50).

Monitoring Policies

You can monitor the policies to see how well they are suited to the zone traffic volume and services.

This section contains the following topics:

Displaying Policies

Displaying Policy Statistics

Displaying Policies

You can display the zone policies to verify that they are adapted to the zone traffic characteristics. You might want to view the zone-constructed policies to verify that these policies are customized for the traffic characteristics of the zone. You can configure only policies that appear in this list.

The Detector module displays only current zone policies. If a policy template was disabled during the policy construction phase, the Detector module does not create policies from that policy template, and you do not see these policies when you enter the show policies command.

To display the zone policies, use the following command in zone configuration mode:

show policies policy-path

The policy-path argument specifies a group of policies. You can use an asterisk (*) as a wildcard character in each policy path section. If you do not specify a policy path section, the Detector module considers the unspecified section to be a wildcard (*). For example, the policy tcp_services//analysis//global uses wildcards for the service and the packet type sections.

To display the statistics of all policies, enter an asterisk (*) for the policy path.

See the "Understanding Zone Policies" section for more information about the policy path sections.

The following example shows how to display all the zone policies:

user@DETECTOR-conf-zone-scannet# show policies *

The following example shows how to display all policies that monitor DNS-over-TCP synchronization packets on port 53:

user@DETECTOR-conf-zone-scannet# show policies dns_tcp/53/*/syns/*

Table 7-18 describes the fields in the show policies command output.

Table 7-18 Field Descriptions of the show policies Command Output 

Field
Description

Policy

Policy name. See the "Understanding Zone Policies" section for more information about the policy path sections.

State

Policy state. See the "Changing the Policy State" section for more information.

act = active, inact = inactive, disab= disabled

IStatus

Policy interactive status. See the "Configuring the Policy Interactive Status" section for more information.

a-accept = always-accept, a-ignor = always-ignore,
interac = interactive

Threshold

Policy threshold. When traffic rate exceeds this threshold, the Detector module executes the action associated with the policy. See the "Configuring the Policy Threshold" section for more information.

List

Number of specific IP thresholds defined for the policy. See the "Configuring Specific IP Thresholds" section for more information. Displays H (histogram) for policies that relate to worms. See the "Understanding Worm Policies" section for more information.

Action

Action that the Detector module executes when the traffic exceeds the policy threshold. See the "Configuring the Policy Action" section for more information.

Timeout

Minimum time span that the policy action is valid. The Detector module determines, according to the filter-termination thresholds, whether or not the dynamic filter that was produced by the policy is to be inactivated. See the "Configuring the Policy Timeout" section for more information.


Displaying Policy Statistics

You can display the rate of the traffic flowing through a zone policy or a group of zone policies and you can determine whether the type of services and volume represent the zone traffic. The Detector module displays the traffic flows forwarded to the zone with the highest rates as measured by the policies. The rate is calculated based on traffic samples.

To display the policy statistics, use the following command in zone configuration mode:

show policies policy-path statistics [num-entries]

Table 7-19 provides the arguments for the show policies statistics command output.

Table 7-19 Arguments for the show policies statistics Command 

Parameter
Description
policy-path

Group of policies for which to display statistics.

You can use an asterisk (*) as a wildcard character in each policy path section. If you do not specify a policy path section, the Detector module relates to the unspecified section as a wildcard (*). For example, the policy tcp_services//analysis//global uses wildcards for the service and the packet type sections.

To display the statistics of all policies, enter an asterisk (*) for the policy-path.

See the "Understanding Zone Policies" section for more information about the policy path sections.

num-entries

(Optional) Number of entries to display. Enter a number from 1 to 100. The Detector module displays the policies with the highest values.


The following example shows how to display the statistics of all the zone policies:

user@DETECTOR-conf-zone-scannet# show policies * statistics

The following example shows how to display the statistics of all policies that monitor DNS-over-TCP synchronization packets on port 53:

user@DETECTOR-conf-zone-scannet# show policies dns_tcp/53/*/syns/*

The following example shows how to display the statistics of the zone global traffic:

user@DETECTOR-conf-zone-scannet# show policies */*/*/*/global statistics 

The Detector module displays the information in four tables. The information in each table is sorted by value, with the highest values appearing at the top of the table.

Table 7-20 displays the fields in the tables in the show policies statistics command output.


Note The Detector module does not display tables that contain no data.


Table 7-20 Field Descriptions of the show policies statistics Command Output Tables 

Column
Description
Fields in all output tables

Key

Key that is the traffic characteristic used to aggregate the policies.

For example, in the tcp_services/any/analysis/syns/dst_ip policy, the key is the destination IP address (dst_ip). If the traffic characteristic that was used to aggregate the policies is global, the key displays N/A.

In policies that relate to worms, such as worm_tcp/any/analysis/non_estb_conns/scanners, the key is the the source IP address that scans the zone network addresses, colon, and the destination port that is being scanned, as shown in this example: 192.128.100.3:70.

See Table 7-8 for more information.

Policy

Policy name. See the "Understanding Zone Policies" section for more information.

Fields in one of the output tables

Rate

Rate of the traffic that flows through the policy and is measured in packets per second. The rate is calculated based on traffic samples.

Connection

Number of concurrent connections. This information is available for tcp_connections policies with a packet type of in_nodata_conns.

Ratio

Ratio between the number of SYN flagged packets and the number of FIN/RST flagged packets. This information is available for syn_by_fin policies only.

Dst IPs

Number of zone destination IP addresses that were scanned. This information is available for worm_tcp policies only.


Backing Up the Policy Configuration

You can back up the current zone policies at any time by using the snapshot threshold-selection cur-thresholds command in zone configuration mode.

The following example shows how to create a snapshot to back up the current policy configuration:

user@DETECTOR-conf-zone-scannet# snapshot threshold-selection cur-thresholds