This chapter describes how Cisco ISE profiles endpoints that connect to a Cisco ISE network, and how you can create profiling policies to manage these endpoints and grant them access to the resources in your network.
The profiling service in Cisco Identity Services Engine (ISE) identifies the devices that connect to your network and their location. The endpoints are profiled based on the endpoint profiling policies configured in Cisco ISE. Cisco ISE then grants permission to the endpoints to access the resources in your network based on the result of the policy evaluation.
The profiling service:
Facilitates an efficient and effective deployment and ongoing management of authentication by using IEEE standard 802.1X port-based authentication access control, MAC Authentication Bypass (MAB) authentication, and Network Admission Control (NAC) for any enterprise network of varying scale and complexity.
Identifies, locates, and determines the capabilities of all of the attached network endpoints regardless of endpoint types.
Protects against inadvertently denying access to some endpoints.
You can use the profiling service to discover, locate, and determine the capabilities of all the endpoints connected to your network. You can ensure and maintain appropriate access of endpoints to the enterprise network, regardless of their device types.
The profiling service collects attributes of endpoints from the network devices and the network, classifies endpoints into a specific group according to their profiles, and stores endpoints with their matched profiles in the Cisco ISE database. All the attributes that are handled by the profiling service need to be defined in the profiler dictionaries.
The profiling service identifies each endpoint on your network, and groups those endpoints according to their profiles to an existing endpoint identity group in the system, or to a new group that you can create in the system. By grouping endpoints, and applying endpoint profiling policies to the endpoint identity group, you can determine the mapping of endpoints to the corresponding endpoint profiling policies.
Cisco ISE Profiler Queue Limit Configuration
Cisco ISE profiler collects a significant amount of endpoint data from the network in a short period of time. It causes Java Virtual Machine (JVM) memory utilization to go up due to accumulated backlog when some of the slower Cisco ISE components process the data generated by the profiler, which results in performance degradation and stability issues.
To ensure that the profiler does not increase the JVM memory utilization and prevent JVM to go out of memory and restart, limits are applied to the following internal components of the profiler:
Endpoint Cache—Internal cache is limited in size that has to be purged periodically (based on least recently used strategy) when the size exceeds the limit.
Forwarder—The main ingress queue of endpoint information collected by the profiler.
Event Handler—An internal queue that disconnects a fast component, which feeds data to a slower processing component (typically related to a database query).
maxEndPointsInLocalDb = 100000 (endpoint objects in cache)
endPointsPurgeIntervalSec = 300 (endpoint cache purge thread interval in seconds)
numberOfProfilingThreads = 8 (number of threads)
The limit is applicable to all profiler internal event handlers. A monitoring alarm is triggered when queue size limit is reached.
NetworkDeviceEventHandler—For network device events, in addition to filtering duplicate Network Access Device (NAD) IP addresses, which are already cached.
ARPCacheEventHandler—For ARP Cache events.
CoAHandler—For CoA events, in addition to filtering duplicate MAC addresses that are still in the queue for processing.
You can configure the profiler queue limits and others as listed below in the platform.properties file. This file is included during Cisco ISE installation, which can be located in the following location: /opt/CSCOcpm/config/
You can configure the profiling service that provides you a contextual inventory of all the endpoints that are using your network resources in any Cisco ISE-enabled network.
You can configure the profiling service to run on a single Cisco ISE node that assumes all Administration, Monitoring, and Policy Service personas by default.
In a distributed deployment, the profiling service runs only on Cisco ISE nodes that assume the Policy Service persona and does not run on other Cisco ISE nodes that assume the Administration and Monitoring personas.
Step 1 Choose Administration > System > Deployment.
Step 2 Choose a Cisco ISE node that assumes the Policy Service persona.
Step 3 Click Edit in the Deployment Nodes page.
Step 4 On the General Settings tab, check the Policy Service check box. If the Policy Service check box is unchecked, both the session services and the profiling service check boxes are disabled.
Step 5 Perform the following tasks:
a. Check the Enable Session Services check box to run the Network Access, Posture, Guest, and Client Provisioning session services.
b. Check the Enable Profiling Services check box to run the profiling service.
Step 6 Click Save to save the node configuration.
Network Probes Used by Profiling Service
Network probe is a method used to collect an attribute or a set of attributes from an endpoint on your network. The probe allows you to create or update endpoints with their matched profile in the Cisco ISE database.
Cisco ISE can profile devices using a number of network probes that analyze the behavior of devices on the network and determine the type of the device. Network probes help you to gain more network visibility.
Cisco ISE profiler implements Cisco IOS NetFlow Version 9. We recommend using NetFlow Version 9, which has additional functionality needed to enhance the profiler to support the Cisco ISE profiling service.
You can collect NetFlow Version 9 attributes from the NetFlow-enabled network access devices to create an endpoint, or update an existing endpoint in the Cisco ISE database. You can configure NetFlow Version 9 to attach the source and destination MAC addresses of endpoints and update them. You can also create a dictionary of NetFlow attributes to support NetFlow-based profiling.
In addition, Cisco ISE supports NetFlow versions earlier than Version 5. If you use NetFlow Version 5 in your network, then you can use Version 5 only on the primary network access device (NAD) at the access layer because it will not work anywhere else.
Cisco IOS NetFlow Version 5 packets do not contain MAC addresses of endpoints. The attributes that are collected from NetFlow Version 5 cannot be directly added to the Cisco ISE database. You can discover endpoints by using their IP addresses, and append the NetFlow Version 5 attributes to endpoints, which can be done by combining IP addresses of the network access devices and IP addresses obtained from the NetFlow Version 5 attributes. However, these endpoints must have been previously discovered with the RADIUS or SNMP probe.
The MAC address is not a part of IP flows in earlier versions of NetFlow Version 5, which requires you to profile endpoints with their IP addresses by correlating the attributes information collected from the network access devices in the endpoints cache.
The Dynamic Host Configuration Protocol probe in your Cisco ISE deployment, when enabled, allows the Cisco ISE profiling service to reprofile endpoints based only on new requests of INIT-REBOOT, and SELECTING message types. Though other DHCP message types such as RENEWING and REBINDING are processed, they are not used for profiling endpoints. Any attribute parsed out of DHCP packets is mapped to endpoint attributes.
DHCPREQUEST Message Generated During INIT-REBOOT State
If the DHCP client checks to verify a previously allocated and cached configuration, then the client must not fill in the Server identifier (server-ip) option. Instead it should fill in the Requested IP address (requested-ip) option with the previously assigned IP address, and fill in the Client IP Address (ciaddr) field with zero in its DHCPREQUEST message. The DHCP server will then send a DHCPNAK message to the client if the Requested IP address is incorrect or the client is located in the wrong network.
DHCPREQUEST Message Generated During SELECTING State
The DHCP client inserts the IP address of the selected DHCP server in the Server identifier (server-ip) option, fills in the Requested IP address (requested-ip) option with the value of the Your IP Address (yiaddr) field from the chosen DHCPOFFER by the client, and fills in the “ciaddr” field with zero.
Table 22-1 describes the different states of DHCP client messages.
Table 22-1 DHCP Client Messages from Different States
Wireless LAN Controller Configuration in DHCP Bridging Mode
We recommend that you configure wireless LAN controllers (WLCs) in Dynamic Host Configuration Protocol (DHCP) bridging mode, where you can forward all the DHCP packets from the wireless clients to Cisco ISE.
You must uncheck the Enable DHCP Proxy check box available in the WLC web interface: Controller > Advanced > DHCP Master Controller Mode > DHCP Parameters.
You must also ensure that the DHCP IP helper command points to the Cisco ISE Policy Service node.
DHCP SPAN Probe
The DHCP Switched Port Analyzer (SPAN) probe, when initialized in a Cisco ISE node, listens to network traffic, which are coming from network access devices on a specific interface. You need to configure network access devices to forward DHCP SPAN packets to the Cisco ISE profiler from the DHCP servers. The profiler receives these DHCP SPAN packets and parses them to capture the attributes of an endpoint, which can be used for profiling endpoints.
In HTTP probe, the identification string is transmitted in an HTTP request-header field User-Agent, which is an attribute that can be used to create a profiling condition of IP type, and to check the web browser information. The profiler captures the web browser information from the User-Agent attribute along with other HTTP attributes from the request messages, and adds them to the list of endpoint attributes.
Cisco ISE listens to communication from the web browsers on both port 80 and port 8080. Cisco ISE provides many default profiles, which are built in to the system to identify endpoints based on the User-Agent attribute.
The HTTP probe in your Cisco ISE deployment, when enabled with the Switched Port Analyzer (SPAN) probe, allows the profiler to capture HTTP packets from the specified interfaces. You can use the SPAN capability on port 80, where the Cisco ISE server listens to communication from the web browsers.
HTTP SPAN collects HTTP attributes of an HTTP request-header message along with the IP addresses in the IP header (L3 header), which can be associated to an endpoint based on the MAC address of an endpoint in the L2 header. This information is useful for identifying different mobile and portable IP-enabled devices such as Apple devices, and computers with different operating systems. Identifying different mobile and portable IP-enabled devices is made more reliable because the Cisco ISE server redirects captures during a guest login or client provisioning download. This allows the profiler to collect the User-Agent attribute and other HTTP attributes, from the request messages and then identify devices such as Apple devices.
Unable to Collect HTTP Attributes in Cisco ISE Running on VMware
If you deploy Cisco ISE on an ESX server (VMware), the Cisco ISE profiler collects the Dynamic Host Configuration Protocol traffic but does not collect the HTTP traffic due to configuration issues on the vSphere client.
To collect HTTP traffic on a VMware setup, configure the security settings by changing the Promiscuous Mode to Accept from Reject (by default) of the virtual switch that you create for the Cisco ISE profiler. When the Switched Port Analyzer (SPAN) probe for DHCP and HTTP is enabled, Cisco ISE profiler collects both the DHCP and HTTP traffic.
You can configure Cisco ISE for authentication with RADIUS, where you can define a shared secret that you can use in client-server transactions. With the RADIUS request and response messages that are received from the RADIUS servers, the profiler can collect RADIUS attributes, which can be used for profiling endpoints.
Cisco ISE can function as a RADIUS server, and a RADIUS proxy client to other RADIUS servers. When it acts as a proxy client, it uses external RADIUS servers to process RADIUS requests and response messages.
The network devices should be configured to send Framed-IP-Address RADIUS attribute #8 to enable the probes to collect the IP address of the endpoints.
Cisco ISE enables you to detect devices in the subnet by using the NMAP manual subnet scan. You must choose the Policy Service node from the Admin portal in your deployment, and run the manual subnet scan from the Policy Service node that is enabled to run the profiling service.
Each NMAP manual subnet scan has a unique numeric ID that is used to update an endpoint source information with that scan ID. Upon detection of endpoints, the endpoint source information can also be updated to indicate that it is discovered by the Network Scan probe.
The NMAP manual subnet scan is useful for detecting devices such as printers with a static IP address assigned to them that are connected constantly to the Cisco ISE network, and therefore these devices cannot be discovered by other probes.
The following are the limitations of a NMAP manual subnet scan:
Scanning a subnet is highly resource intensive.
Scanning a subnet is lengthy process that depends on the size and density of the subnet.
Number of active scans is always restricted to one scan, which means that you can scan only a single subnet at a time.
You can cancel a subnet scan at any time while the subnet scan is in progress. You can use the Click to see latest scan results link to view the most recent network scan results that are stored in Administration > Identities > Latest Network Scan Results.
Scans only specified ports. For example, U:161, 162
Network Mapper Integration with Cisco ISE
Network Mapper (NMAP) is integrated with the Cisco ISE profiler to augment the profiling service capability to improve endpoint classification, particularly iDevices and other mobile devices. You can either perform a manual subnet scan on a specific subnet by using the Network Scan probe, or you can associate a network scan action to an endpoint profile (a specific profile) to perform a scan on an endpoint.
SNMP Read Only Community Strings for NMAP Manual Subnet Scan
The NMAP manual subnet scan is augmented with an SNMP Query whenever the scan discovers that UDP port 161 is open on an endpoint that results in more attributes being collected. During the NMAP manual subnet scan, the Network Scan probe detects whether SNMP port 161 is open on the device. If the port is open, an SNMP Query is triggered with a default community string (public). If the device supports SNMP and the default Read Only community string is set to public, you can obtain the MAC address of the device from the MIB value “ifPhysAddress”.
In addition, you can configure additional SNMP Read Only community strings separated by a comma for the NMAP manual network scan in the Profiler Configuration page. You can also specify new Read Only community strings for an SNMP MIB walk with SNMP versions 1 and 2c in the following location: Administration > System > Settings > Profiling.
You can create or update endpoints only by using their MAC addresses in an enterprise network. If you do not find an entry in the ARP cache, then you can create or update endpoints by using the L2 MAC address of an HTTP packet and the IN_SRC_MAC of a NetFlow packet in Cisco ISE.
The profiling service is dependent on L2 adjacency when endpoints are only a hop away. When endpoints are L2 adjacent, the IP addresses and MAC addresses of endpoints are already mapped, and there is no need for IP-MAC cache mapping. If endpoints are not L2 adjacent and are multiple hops away, mapping may not be reliable.
Some of the known attributes of NetFlow packets that you collect include PROTOCOL, L4_SRC_PORT, IPV4_SRC_ADDR, L4_DST_PORT, IPV4_DST_ADDR, IN_SRC_MAC, OUT_DST_MAC, IN_SRC_MAC, and OUT_SRC_MAC. When endpoints are not L2 adjacent and are multiple L3 hops away, the IN_SRC_MAC attributes carry only the MAC addresses of L3 network devices.
When the HTTP probe is enabled in Cisco ISE, you can create endpoints only by using the MAC addresses of HTTP packets, because the HTTP request messages do not carry IP addresses and MAC addresses of endpoints in the payload data.
Cisco ISE implements an ARP cache in the profiling service, so that you can reliably map the IP addresses and the MAC addresses of endpoints. For the ARP cache to function, you must enable either the DHCP probe or the RADIUS probe. The DHCP and RADIUS probes carry the IP addresses and the MAC addresses of endpoints in the payload data. The dhcp-requested address attribute in the DHCP probe and the Framed-IP-address attribute in the RADIUS probe carry the IP addresses of endpoints, along with their MAC addresses, which can be mapped and stored in the ARP cache.
The Domain Name Service (DNS) probe in your Cisco ISE deployment allows the profiler to lookup an endpoint and get the fully qualified domain name (FQDN). After an endpoint is detected in your Cisco ISE-enabled network, a list of endpoint attributes is collected from the NetFlow, DHCP, DHCP SPAN, HTTP, RADIUS, or SNMP probes.
When you deploy Cisco ISE in a standalone or in a distributed environment for the first time, you are prompted to run the setup utility to configure the Cisco ISE appliance. When you run the setup utility, you will configure the Domain Name System (DNS) domain and the primary nameserver (primary DNS server), where you can configure one or more nameservers during setup. You can also change or add DNS nameservers later after deploying Cisco ISE using the CLI commands.
Before a DNS lookup can be performed, one of the following probes must be started along with the DNS probe: DHCP, DHCP SPAN, HTTP, RADIUS, or SNMP. This allows the DNS probe in the profiler to do a reverse DNS lookup (FQDN lookup) against specified name servers that you define in your Cisco ISE deployment. A new attribute is added to the attribute list for an endpoint, which can be used for an endpoint profiling policy evaluation. The FQDN is the new attribute that exists in the system IP dictionary. You can create an endpoint profiling condition to validate the FQDN attribute and its value for profiling.
The following are the specific endpoint attributes that are required for a DNS lookup and the probe that collects these attributes:
The dhcp-requested-address attribute—An attribute collected by the DHCP and DHCP SPAN probes
The SourceIP attribute—An attribute collected by the HTTP probe
The Framed-IP-Address attribute—An attribute collected by the RADIUS probe
The cdpCacheAddress attribute—An attribute collected by the SNMP probe
DNS Lookup with an Inline Posture Node Deployment in Bridged Mode
For the Domain Name Service probe to work with Inline Posture deployment in the Bridged mode, you must configure the callStationIdType information sent in RADIUS messages for the Wireless LAN Controllers (WLCs).
The Framed-IP-Address attribute in RADIUS messages does not contain the Call Station ID type in the MAC address format. Therefore RADIUS messages cannot be associated with the MAC address of endpoints, and the DNS probe is unable to perform the reverse DNS lookup. In order to profile endpoints, you must enable the RADIUS, and DNS probes in Cisco ISE, and then configure the WLCs to send the calling station ID in the MAC address format instead of the current IP address format in RADIUS messages.
The WLCs must be configured to send the calling station ID in the MAC address format instead of the current IP address format in RADIUS messages.
Once the callStationIdType is configured in the WLCs, the configuration uses the selected calling station ID for communications with RADIUS servers and other applications. It results in endpoints authentication, and then the DNS probe does a reverse DNS lookup (FQDN lookup) against the specified name servers and update the FQDN of endpoints.
Configuring Call Station ID Type in the WLC Web Interface
You can use the WLC web interface to configure Call Station ID Type information. You can go to the Security tab of the WLC web interface to configure the calling station ID in the RADIUS Authentication Servers page. The MAC Delimiter field is set to Colon by default in the WLC user interface.
Step 1 Log in to your Wireless LAN Controller user interface.
Step 2 Click Security.
Step 3 Expand AAA, and then choose RADIUS > Authentication.
Step 4 Choose System MAC Address from the Call Station ID Type drop-down list.
Step 5 Check the AES Key Wrap check box when you run Cisco ISE in FIPS mode.
Step 6 Choose Colon from the MAC Delimeter drop-down list.
In addition to configuring the SNMP Query probe in the Edit Node page, you must configure other Simple Management Protocol settings in the following location: Administration > Network Resources > Network Devices.
You can configure SNMP settings in the new network access devices (NADs) in the Network Devices list page. The polling interval that you specify in the SNMP query probe or in the SNMP settings in the network access devices query NADs at regular intervals.
You can turn on and turn off SNMP querying for specific NADs based on the following configurations:
SNMP query on Link up and New MAC notification turned on or turned off
SNMP query on Link up and New MAC notification turned on or turned off for Cisco Discovery Protocol information
SNMP query timer for once an hour for each switch by default
For an iDevice, and other mobile devices that do not support SNMP, the MAC address can be discovered by the ARP table, which can be queried from the network access device by an SNMP Query probe.
When you configure SNMP settings on the network devices, you must ensure that the Cisco Discovery Protocol is enabled (by default) on all the ports of the network devices. If you disable the Cisco Discovery Protocol on any of the ports on the network devices, then you may not be able to profile properly because you will miss the Cisco Discovery Protocol information of all the connected endpoints.
You can enable the Cisco Discovery Protocol globally by using the cdp run command on a network device, and enable the Cisco Discovery Protocol by using the cdp enable command on any interface of the network access device. To disable the Cisco Discovery Protocol on the network device and on the interface, use the no keyword at the beginning of the commands.
Link Layer Discovery Protocol Support with SNMP Query
The Cisco ISE profiler uses an SNMP Query to collect LLDP attributes. You can also collect LLDP attributes from a Cisco IOS sensor, which is embedded in the network device, by using the RADIUS probe.
Table 22-3 shows the default LLDP configuration settings that you can use to configure LLDP global configuration and LLDP interface configuration commands on the network access devices.
Table 22-3 Default LLDP Configuration
LLDP global state
LLDP holdtime (before discarding)
LLDP timer (packet update frequency)
LLDP reinitialization delay
Enabled to send and receive all TLVs.
LLDP interface state
Enabled to send all LLDP-MED TLVs
CDP and LLDP Capability Codes Displayed in a Single Character
The Attribute List of an endpoint displays a single character value for the lldpCacheCapabilities and lldpCapabilitiesMapSupported attributes. The values are the Capability Codes that are displayed for the network access device that runs CDP and LLDP.
R - Router, T - Trans Bridge, B - Source Route Bridge, S - Switch, H - Host, I - IGMP,
r - Repeater, P - Phone, D - Remote, C - CVTA, M - Two-port Mac Relay
(W) WLAN Access Point, (P) Repeater, (S) Station, (O) Other
SNMP Trap Probe
The SNMP Trap receives information from the specific network access devices that support MAC notification, linkup, linkdown, and informs. The SNMP Trap probe receives information from the specific network access devices when ports come up or go down and endpoints disconnect from or connect to your network, which results in the information received that is not sufficient to create endpoints in Cisco ISE.
For SNMP Trap to be fully functional and create endpoints, you must enable SNMP Query so that the SNMP Query probe triggers a poll event on the particular port of the network access device when a trap is received. To make this feature fully functional you should configure the network access device and SNMP Trap.
Note Cisco ISE does not support SNMP Traps that are received from the Wireless LAN Controllers (WLCs) and Access Points (APs).
Setting up COA, SNMP RO Community and Endpoint Attribute Filter
Cisco ISE allows a global configuration to issue a Change of Authorization (CoA) in the Profiler Configuration page that enables the profiling service with more control over endpoints that are already authenticated.
In addition, you can configure additional SNMP Read Only community strings separated by a comma for the NMAP manual network scan in the Profiler Configuration page. The SNMP RO community strings are used in the same order as they appear in the Current custom SNMP community strings field.
You can also configure endpoint attribute filtering in the Profiler Configuration page.
Step 1 Choose Administration > System > Settings > Profiling.
Step 2 Choose one of the following settings to configure the CoA type:
No CoA (default)—You can use this option to disable the global configuration of CoA. This setting overrides any configured CoA per endpoint profiling policy.
Port Bounce —You can use this option, if the switch port exists with only one session. If the port exists with multiple sessions, then use the Reauth option.
Reauth —You can use this option to enforce reauthentication of an already authenticated endpoint when it is profiled.
If you have multiple active sessions on a single port, the profiling service issues a CoA with the Reauth option even though you have configured CoA with the Port Bounce option. This function avoids disconnecting other sessions, a situation that might occur with the Port Bounce option.
Step 3 Enter new SNMP community strings separated by a comma for the NMAP manual network scan in the Change custom SNMP community strings field, and re-enter the strings in the Confirm custom SNMP community strings field for confirmation.
Step 4 Check the Endpoint Attribute Filter check box to enable endpoint attribute filtering.
Global Configuration of Change of Authorization for Authenticated Endpoints
You can use the global configuration option to disable change of authorization (CoA) by using the default No CoA option or enable CoA by using port bounce and reauthentication options. If you have configured Port Bounce for CoA in Cisco ISE, the profiling service may still issue other CoAs as described in the “CoA Exemptions” section.
You can use the RADIUS probe or the Monitoring persona REST API to authenticate the endpoints. You can enable the RADIUS probe, which allows faster performance. If you have enabled CoA, then we recommend that you enable the RADIUS probe in conjunction with your CoA configuration in the Cisco ISE application for faster performance. The profiling service can then issue an appropriate CoA for endpoints by using the RADIUS attributes that are collected.
If you have disabled the RADIUS probe in the Cisco ISE application, then you can rely on the Monitoring persona REST API to issue CoAs. This allows the profiling service to support a wider range of endpoints. In a distributed deployment, your network must have at least one Cisco ISE node that assumes the Monitoring persona to rely on the Monitoring persona REST API to issue a CoA.
Cisco ISE arbitrarily will designate either the primary or secondary Monitoring node as the default destination for REST queries in your distributed deployment, because both the primary and secondary Monitoring nodes have identical session directory information.
The profiling service issues the change of authorization in the following cases:
Endpoint deleted—When an endpoint is deleted from the Endpoints page and the endpoint is disconnected or removed from the network.
An exception action is configured—If you have an exception action configured per profile that leads to an unusual or an unacceptable event from that endpoint. The profiling service moves the endpoint to the corresponding static profile by issuing a CoA.
An endpoint is profiled for the first time—When an endpoint is not statically assigned and profiled for the first time; for example, the profile changes from an unknown to a known profile.
An endpoint identity group has changed—When an endpoint is added or removed from an endpoint identity group that is used by an authorization policy.
The profiling service issues a CoA when there is any change in an endpoint identity group, and the endpoint identity group is used in the authorization policy for the following:
– The endpoint identity group changes for endpoints when they are dynamically profiled
– The endpoint identity group changes when the static assignment flag is set to true for a dynamic endpoint
An endpoint profiling policy has changed and the policy is used in an authorization policy—When an endpoint profiling policy changes, and the policy is included in a logical profile that is used in an authorization policy. The endpoint profiling policy may change due to the profiling policy match or when an endpoint is statically assigned to an endpoint profiling policy, which is associated to a logical profile. In both the cases, the profiling service issues a CoA, only when the endpoint profiling policy is used in an authorization policy.
The profiling service does not issue a CoA when there is a change in an endpoint identity group and the static assignment is already true.
Cisco ISE does not issue a CoA for the following reasons:
An Endpoint disconnected from the network—When an endpoint disconnected from your network is discovered.
Authenticated wired (Extensible Authentication Protocol) EAP-capable endpoint—When an authenticated wired EAP-capable endpoint is discovered.
Multiple active sessions per port—When you have multiple active sessions on a single port, the profiling service issues a CoA with the Reauth option even though you have configured CoA with the Port Bounce option.
Packet-of-Disconnect CoA (Terminate Session) when a wireless endpoint is detected—If an endpoint is discovered as wireless, then a Packet-of-Disconnect CoA (Terminate-Session) is issued instead of the Port Bounce CoA. The benefit of this change is to support the Wireless LAN Controller (WLC) CoA.
An Endpoint Created through Guest Device Registration flow—When endpoints are created through device registration for the guests. Even though CoA is enabled globally in Cisco ISE, the profiling service does not issue a CoA so that the device registration flow is not affected. In particular, the PortBounce CoA global configuration breaks the flow of the connecting endpoint.
Global No CoA Setting overrides Policy CoA—Global No CoA overrides all configuration settings in endpoint profiling policies as there is no CoA issued in Cisco ISE irrespective of CoA configured per endpoint profiling policy.
Note No CoA and Reauth CoA configurations are not affected, and the profiler service applies the same CoA configuration for wired and wireless endpoints.
Change of Authorization Issued for Each Type of CoA Configuration
Table 22-4 summarizes CoA for different environments for each type of CoA configuration in Cisco ISE.
Table 22-4 Change of Authorization for Each CoA Configuration
No CoA Configuration
Port Bounce Configuration
Global CoA configuration in Cisco ISE (typical configuration)
An endpoint is disconnected on your network
Change of authorization is determined by the RADIUS attribute Acct-Status -Type value Stop.
Wired with multiple active sessions on the same switch port
Reauthentication avoids disconnecting other sessions.
Packet-of-Disconnect CoA (Terminate Session)
Support to Wireless LAN Controller.
Incomplete CoA data
Due to missing RADIUS attributes.
Attribute Filters for ISE Database Persistence and Performance
Cisco ISE implements filters for Dynamic Host Configuration Protocol (both DHCP Helper and DHCP SPAN), HTTP, RADIUS, and Simple Network Management Protocol probes except for the NetFlow probe to address performance degradation. Each probe filter contains the list of attributes that are temporal and irrelevant for endpoint profiling and removes those attributes from the attributes collected by the probes.
The isebootstrap log (isebootstrap-yyyymmdd-xxxxxx.log) contains messages that handles the creation of dictionaries and with filtering of attributes from the dictionaries. You can also configure to log a debug message when endpoints go through the filtering phase to indicate that filtering has occurred.
The Cisco ISE profiler invokes the following endpoint attribute filters:
A DHCP filter for both the DHCP Helper and DHCP SPAN contains all the attributes that are not necessary and they are removed after parsing DHCP packets. The attributes after filtering are merged with existing attributes in the endpoint cache for an endpoint.
An HTTP filter is used for filtering attributes from HTTP packets, where there is no significant change in the set of attributes after filtering.
A RADIUS filter is used once the syslog parsing is complete and endpoint attributes are merged into the endpoint cache for profiling.
SNMP filter for SNMP Query includes separate CDP and LLDP filters, which are all used for SNMP-Query probe.
Global Setting to Filter Endpoint Attributes with Whitelist
You can reduce the number of persistence events and replication events by reducing the number of endpoint attributes that do not change frequently at the collection point. Enabling the EndPoint Attribute Filter will have the Cisco ISE profiler only to keep significant attributes and discard all other attributes. Significant attributes are those used by the Cisco ISE system or those used specifically in a endpoint profiling policy or rule.
A whitelist is a set of attributes that are used in custom endpoint profiling policies for profiling endpoints, and that are essential for Change of Authorization (CoA), Bring Your Own Device (BYOD), Device Registration WebAuth (DRW), and so on to function in Cisco ISE as expected. The whitelist is always used as a criteria when ownership changes for the endpoint (when attributes are collected by multiple Policy Service nodes) even when disabled.
By default, the whitelist is disabled and the attributes are dropped only when the attribute filter is enabled. The white list is dynamically updated when endpoint profiling policies change including from the feed to include new attributes in the profiling policies. Any attribute that is not present in the whitelist is dropped immediately at the time of collection, and the attribute cannot participate in profiling endpoints. When combined with the buffering, the number of persistence events can be reduced.
You must ensure that the whitelist contains a set of attributes determined from the following two sources:
A set of attributes that are used in the default profiles so that you can match endpoints to the profiles.
A set of attributes that are essential for Change of Authorization (CoA), Bring Your Own Device (BYOD), Device Registration WebAuth (DRW), and so on to function as expected.
Table 22-5 Whitelist Attributes
Certificate Expiration Date
Certificate Issue Date
Certificate Issuer Name
Certificate Serial Number
Total Certainty Factor
Attributes Collection from IOS Sensor Embedded Switches
An IOS sensor integration allows Cisco ISE run time and the Cisco ISE profiler to collect any or all of the attributes that are sent from the switch. You can collect DHCP, CDP, and LLDP attributes directly from the switch by using the RADIUS protocol. The attributes that are collected for DHCP, CDP, and LLDP are then parsed and mapped to attributes in the profiler dictionaries in the following location: Policy > Policy Elements > Dictionaries.
You must enable the network access devices to add IOS sensor protocol data to the RADIUS accounting messages and to generate additional accounting events when it detects new sensor protocol data. This means that any RADIUS accounting message should include all CDP, LLDP, and DHCP attributes.
Enter the following global command:
– Disabling Accounting Augmentation
To disable (accounting) network access devices and add IOS sensor protocol data to the RADIUS accounting messages for sessions that are hosted on a given port (if the accounting feature is globally enabled), enter the following command at the appropriate port:
no device-sensor accounting
– TLV Change Tracking
By default, for each supported peer protocol, client notifications and accounting events are generated only when an incoming packet includes a type, length, and value (TLV) that has not been received previously in the context of a given session.
You must enable client notifications and accounting events for all TLV changes where there are either new TLVs, or where previously received TLVs have different values. Enter the following command:
device-sensor notify all-changes
Be sure that you disable the IOS Device Classifier (local analyzer) in the network access devices.
Enter the following command:
no macro auto monitor
Note This command prevents network access devices from sending two identical RADIUS accounting messages per change.
Endpoint Profiling Policy Rules
You can define a rule that allows you to choose one or more profiling conditions from the library that are previously created and saved in the policy elements library, and to associate an integer value for the certainty factor for each condition, or associate either an exception action or a network scan action for that condition. The exception action or the network scan action is used to trigger the configurable action while Cisco ISE is evaluating the profiling policies with respect to the overall classification of endpoints.
When the rules in a given policy are evaluated separately with an OR operator, the certainty metric for each rule contributes to the overall matching of the endpoint profiles into a specific category of endpoints. If the rules of an endpoint profiling policy match, then the profiling policy and the matched policy are the same for that endpoint when they are dynamically discovered on your network.
Logically Grouped Conditions in Rules
An endpoint profiling policy (profile) contains a single condition or a combination of multiple single conditions that are logically combined using an AND or OR operator, against which you can check, categorize, and group endpoints for a given rule in a policy.
A condition is used to check the collected endpoint attribute value against the value specified in the condition for an endpoint. If you map more than one attribute, you can logically group the conditions, which helps you to categorize endpoints on your network. You can check endpoints against one or more such conditions with a corresponding certainty metric (an integer value that you define) associated with it in a rule or trigger an exception action that is associated to the condition or a network scan action that is associated to the condition.
When you import endpoint profiling policies in Cisco ISE 1.2 from Cisco ISE 1.1.x, you must modify the name of the conditions in the exported profilier_policies.xml before importing policies.
The minimum certainty metric in the profiling policy evaluates the matching profile for an endpoint. Each rule in an endpoint profiling policy has a minimum certainty metric (an integer value) associated to the profiling conditions. The certainty metric is a measure that is added for all the valid rules in an endpoint profiling policy, which measures how each condition in an endpoint profiling policy contributes to improve the overall classification of endpoints.
The certainty metric for each rule contributes to the overall matching of the endpoint profiles into a specific category of endpoints. The certainty metric for all the valid rules are added together to form the matching certainty. It must exceed the minimum certainty factor that is defined in an endpoint profiling policy. By default, the minimum certainty factor for all new profiling policy rules and predefined profiling policies is 10.
Creating Endpoint Profiling Policies
You can use the Profiling Policies page to manage endpoint profiling policies that you create as an administrator of Cisco ISE, and also endpoint profiling policies that are provided by Cisco ISE when deployed.
You can create new profiling policies to profile endpoints by using the following options in the New Profiler Policy page:
Create an Identity Group for the policy to create a matching endpoint identity group or use the endpoint identity group hierarchy
Associated CoA Type
Note When you choose to create an endpoint policy in the Profiling Policies page, do not use the Stop button on your web browsers. This action leads to the following: stops loading the New Profiler Policy page, loads other list pages and the menus within the list pages when you access them, and prevents you from performing operations on all the menus within the list pages except the Filter menus. You might need to log out of Cisco ISE, and then log in again to perform operations on all the menus within the list pages.
You can create a similar characteristic profiling policy by duplicating an endpoint profiling policy through which you can modify an existing profiling policy instead of creating a new profiling policy by redefining all conditions.
Step 3 Enter a name and description for the new endpoint policy that you want to create. The Policy Enabled check box is checked by default to include the endpoint profiling policy for validation when you profile an endpoint.
Step 4 Enter a value for the minimum certainty factor within the valid range 1 to 65535.
Step 5 Click the arrow next to the Exception Action drop-down list to associate an exception action or click the arrow next to the Network Scan (NMAP) Action drop-down list to associate a network scan action.
Step 6 Choose one of the following options for Create an Identity Group for the policy :
Yes, create matching Identity Group
No, use existing Identity Group hierarchy
Step 7 Click the arrow next to the Parent Policy drop-down list to associate a parent policy to the new endpoint policy.
Step 8 Choose a CoA type to be associated in the Associated CoA Type drop-down list.
Step 9 Click in the rule to add conditions and associate an integer value for the certainty factor for each condition or associate either an exception action or a network scan action for that condition for the overall classification of an endpoint.
Step 10 Click Submit to add an endpoint policy or click the Profiler Policy List link from the New Profiler Policy page to return to the Profiling Policies page.
Change of Authorization Configuration per Endpoint Profiling Policy
In addition to the global configuration of change of authorization (CoA) types in Cisco ISE, you can also configure to issue a specific type of CoA associated for each endpoint profiling policy.
The global No CoA type configuration overrides each CoA type configured in an endpoint profiling policy. If the global CoA type is set other than the No CoA type, then each endpoint profiling policy is allowed to override the global CoA configuration.
When a CoA is triggered, each endpoint profiling policy can determine the actual CoA type, as follows:
General Setting—This is the default setting for all the endpoint profiling policies that issues a CoA per global configuration.
No CoA—This setting overrides any global configuration and disables CoA for the profile.
Port Bounce—This setting overrides the global Port Bounce and Reauth configuration types, and issues port bounce CoA.
Reauth—This setting overrides the global Port Bounce and Reauth configuration types, and issues reauthentication CoA.
Note If the profiler global CoA configuration is set to Port Bounce (or Reauth), ensure that you configure corresponding endpoint profiling policies with No CoA, the per-policy CoA option so that the BYOD flow does not break for your mobile devices.
Table 22-6 describes configuration combined for all the CoA types and the actual CoA type issued in each case based on the global and endpoint profiling policy settings.
Table 22-6 CoA Type Issued for Various Combination of Configuration
Global CoA Type
Default CoA Type set per Policy
No coA Type per Policy
Port Bounce Type per Policy
Reauth Type per Policy
Importing Endpoint Profiling Policies
You can import endpoint profiling policies from a file in XML by using the same format that you can create in the export function. If you import newly created profiling policies that have parent policies associated, then you must have defined parent policies before you define child policies.
The imported file contains the hierarchy of endpoint profiling policies that contain the parent policy first, then the profile that you imported next along with the rules and checks that are defined in the policy.
You can export endpoint profiling policies to other Cisco ISE deployments. Or, you can use the XML file as a template for creating your own policies to import. You can also download the file to your system in the default location, which can be used for importing later.
A dialog appears when you want to export endpoint profiling policies, which prompts you to open the profiler_policies.xml with an appropriate application or save it. This is a file in XML format that you can open in a web browser, or in other appropriate applications.
Cisco ISE includes predefined default profiling policies when Cisco ISE is deployed, and their hierarchical construction allows you to categorize identified endpoints on your network, and assign them to a matching endpoint identity groups. Because endpoint profiling policies are hierarchical, you can find that the Profiling Policies page displays the list of generic (parent) policies for devices and child policies to which their parent policies are associated in the Profiling Policies list page.
The Profiling Policies page displays endpoint profiling policies with their names, type, description and the status, if enabled or not for validation.
The endpoint profiling policy types are classified as follows:
Cisco Provided—Endpoint profiling policies that are predefined in Cisco ISE are identified as the Cisco Provided type.
Administrator Modified—Endpoint profiling policies are identified as the Administrator Modified type when you modify predefined endpoint profiling policies. Cisco ISE overwrites changes that you have made in the predefined endpoint profiling policies during upgrade.
You can delete administrator-modified policies but Cisco ISE replaces them with up-to-date versions of Cisco-provided policies.
Administrator Created—Endpoint profiling policies that you create or when you duplicate Cisco-provided endpoint profiling policies are identified as the Administrator Created type.
We recommend that you create a generic policy (a parent) for a set of endpoints from which its children can inherit the rules and conditions. If an endpoint has to be classified, then the endpoint profile has to first match the parent, and then its descendant (child) policies when you are profiling an endpoint.
For example, Cisco-Device is a generic endpoint profiling policy for all Cisco devices, and other policies for Cisco devices are children of Cisco-Device. If an endpoint has to be classified as a Cisco-IP-Phone 7960, then the endpoint profile for this endpoint has to first match the parent Cisco-Device policy, its child Cisco-IP-Phone policy, and then the Cisco-IP-Phone 7960 profiling policy for better classification.
Predefined Endpoint Profiling Policies Overwritten During Upgrade
You can edit existing endpoint profiling policies in the Profiling Policies page. You must also save all your configurations in a copy of the predefined endpoint profiles when you want to modify the predefined endpoint profiling policies.
During an upgrade, Cisco ISE overwrites any configuration that you have saved in the predefined endpoint profiles.
Unable to Delete Endpoint Profiling Policies
You can delete selected or all the endpoint profiling policies in the Profiling Policies page. By default, you can delete all the endpoint profiling policies from the Profiling Policies page. When you select all the endpoint profiling policies and try to delete them in the Profiling Policies page, some of them may not be deleted when the endpoint profiling policies are a parent policy mapped to other endpoint profiling policies or mapped to an authorization policy and a parent policy to other endpoint profiling policies.
1. You cannot delete Cisco Provided endpoint profiling policies,
1. You cannot delete a parent profile in the Profiling Policies page when an endpoint profile is defined as a parent to other endpoint profiles. For example, Cisco-Device is a parent to other endpoint profiling policies for Cisco devices.
2. You cannot delete an endpoint profile when it is mapped to an authorization policy. For example, Cisco-IP-Phone is mapped to the Profiled Cisco IP Phones authorization policy, and it is a parent to other endpoint profiling policies for Cisco IP Phones.
Predefined Profiling Policies for Draeger Medical Devices
Cisco ISE contains default endpoint profiling policies that include a generic policy for Draeger medical devices, a policy for Draeger-Delta medical device, and a policy for Draeger-M300 medical device. Both the medical devices share ports 2050 and 2150, and therefore you cannot classify the Draeger-Delta and Draeger-M300 medical devices when you are using the default Draeger endpoint profiling policies.
If these Draeger devices share ports 2050 and 2150 in your environment, you must add a rule in addition to checking for the device destination IP address in the default Draeger-Delta and Draeger-M300 endpoint profiling policies so that you can distinquish these medical devices.
Cisco ISE includes the following profiling conditions that are used in the endpoint profiling policies for the Draeger medical devices:
Draeger-Delta-PortCheck1 that contains port 2000
Draeger-Delta-PortCheck2 that contains port 2050
Draeger-Delta-PortCheck3 that contains port 2100
Draeger-Delta-PortCheck4 that contains port 2150
Draeger-M300PortCheck1 that contains port 1950
Draeger-M300PortCheck2 that contains port 2050
Draeger-M300PortCheck3 that contains port 2150
Endpoint Profiling Policy for Unknown Endpoints
An endpoint that does not match existing profiles and cannot be profiled in Cisco ISE is an unknown endpoint. An unknown profile is the default system profiling policy that is assigned to an endpoint, where an attribute or a set of attributes collected for that endpoint do not match with existing profiles in Cisco ISE.
An Unknown profile is assigned in the following scenarios:
When an endpoint is dynamically discovered in Cisco ISE, and there is no matching endpoint profiling policy for that endpoint, it is assigned to the unknown profile.
When an endpoint is statically added in Cisco ISE, and there is no matching endpoint profiling policy for a statically added endpoint, it is assigned to the unknown profile.
If you have statically added an endpoint to your network, the statically added endpoint is not profiled by the profiling service in Cisco ISE. You can change the unknown profile later to an appropriate profile and Cisco ISE will not reassign the profiling policy that you have assigned.
Endpoint Profiling Policy for Statically Added Endpoints
For the endpoint that is statically added to be profiled, the profiling service computes a profile for the endpoint by adding a new MATCHEDPROFILE attribute to the endpoint. The computed profile is the actual profile of an endpoint if that endpoint is dynamically profiled. This allows you to find the mismatch between the computed profile for statically added endpoints and the matching profile for dynamically profiled endpoints.
Endpoint Profiling Policy for Static IP Devices
If you have an endpoint with a statically assigned IP address, you can create a profile for such static IP devices.
You must enable the RADIUS probe or SNMP Query and SNMP Trap probes to profile an endpoint that has a static IP address.
Endpoint Profiling Policy Matching
Cisco ISE always considers a chosen policy for an endpoint that is the matched policy rather than an evaluated policy when the profiling conditions that are defined in one or more rules are met in a profiling policy. Here, the status of static assignment for that endpoint is set to false in the system. But, this can be set to true after it is statically reassigned to an existing profiling policy in the system, by using the static assignment feature during an endpoint editing.
The following apply to the matched policies of endpoints:
For statically assigned endpoint, the profiling service computes the MATCHEDPROFILE.
For dynamically assigned endpoints, the MATCHEDPROFILEs are identical to the matching endpoint profiles.
You can determine a matching profiling policy for dynamic endpoints using one or more rules that are defined in a profiling policy and assign appropriately an endpoint identity group for categorization.
When an endpoint is mapped to an existing policy, the profiling service searches the hierarchy of profiling policies for the closest parent profile that has a matching group of policies and assigns the endpoint to the appropriate endpoint policy.
Endpoint Profiling Policies Used for Authorization
You can use an endpoint profiling policy in authorization rules, where you can create a new condition to include a check for an endpoint profiling policy as an attribute, and the attribute value assumes the name of the endpoint profiling policy. You can select an endpoint profiling policy from the EndPoints dictionary, which includes the following attributes: PostureApplicable, EndPointPolicy, LogicalProfile, and BYODRegistration.
You can define an authorization rule that includes a combination of EndPointPolicy, BYODRegistration, and identity groups.
Endpoint Profiling Policies Grouped into Logical Profiles
A logical profile is a container for a category of profiles or associated profiles, irrespective of Cisco-provided or administrator-created endpoint profiling policies. An endpoint profiling policy can be associated to multiple logical profiles.
You can use the logical profile in an authorization policy condition to help create an overall network access policy for a category of profiles. You can create a simple condition for authorization, which can be included in the authorization rule. The attribute-value pair that you can use in the authorization condition is the logical profile (attribute) and the name of the logical profile (value), which can be found in the EndPoints systems dictionary.
For example, you can create a logical profile for all mobile devices like Android, Apple iPhone, or Blackberry by assigning matching endpoint profiling policies for that category to the logical profile. Cisco ISE contains IP-Phone, a default logical profile for all the IP phones, which includes IP-Phone, Cisco-IP-Phone, Nortel-IP-Phone-2000-Series, and Avaya-IP-Phone profiles.
You can create a logical profile that you can use to group a category of endpoint profiling policies, which allows you to create an overall category of profiles or associated profiles. You can also remove the endpoint profiling policies from the assigned set moving them back to the available set.
An exception action is a single configurable action that can be referred to in an endpoint profiling policy, and that is triggered when the exception conditions that are associated with the action are met.
Exception Actions can be any one of the following types:
Cisco-provided—You can not delete Cisco-provided exception actions. Cisco ISE triggers the following noneditable profiling exception actions from the system when you want to profile endpoints in Cisco ISE:
– Authorization Change—The profiling service issues a change of authorization when an endpoint is added or removed from an endpoint identity group that is used by an authorization policy.
– Endpoint Delete—An exception action is triggered in Cisco ISE and a CoA is issued when an endpoint is deleted from the system in the Endpoints page, or reassigned to the unknown profile from the edit page on a Cisco ISE network.
– FirstTimeProfiled—An exception action is triggered in Cisco ISE and a CoA is issued when an endpoint is profiled in Cisco ISE for the first time, where the profile of that endpoint changes from an unknown profile to an existing profile but that endpoint is not successfully authenticated on a Cisco ISE network.
Administrator-created—Cisco ISE triggers profiling exception actions that you create.
You can define and associate one or more exception rules to a single profiling policy. This association triggers an exception action (a single configurable action) when the profiling policy matches and at least one of the exception rules matches in the profiling endpoints in Cisco ISE.
An endpoint scan action is a configurable action that can be referred to in an endpoint profiling policy, and that is triggered when the conditions that are associated with the network scan action are met.
An endpoint scan is used to scan endpoints in order to limit resources usage in the Cisco ISE system. A network scan action scans a single endpoint, unlike resource-intensive network scans. It improves the overall classification of endpoints, and redefines an endpoint profile for an endpoint. Endpoint scans can be processed only one at a time.
You can associate a single network scan action to an endpoint profiling policy. Cisco ISE predefines three scanning types for a network scan action, which can include one or all three scanning types: for instance, an OS-scan, an SNMPPortsAndOS-scan, and a CommonPortsAndOS-scan. You cannot edit or delete OS-scan, SNMPPortsAndOS-scan, and CommonPortsAndOS-scans, which are predefined network scan actions in Cisco ISE. You can also create a new network scan action of your own.
Once an endpoint is appropriately profiled, the configured network scan action cannot be used against that endpoint. For example, scanning an Apple-Device allows you to classify the scanned endpoint to an Apple device. Once an OS-scan determines the operating system that an endpoint is running, it is no longer matched to an Apple-Device profile, but it is matched to an appropriate profile for an Apple device.
A network scan action that is associated with an endpoint profiling policy scans an endpoint for an operating system, Simple Network Management Protocol (SNMP) ports, and common ports.
You can define and associate one or more network scan rules in a endpoint profiling policy. You can also define the type of scanning in each network scan actions. This association triggers a network scan action when the profiling policy matches and at least one of the network scan rules matches in profiling endpoints in Cisco ISE.
Before You Begin
You must ensure that the Network Scan (NMAP) probe is enabled to trigger a network scan action that you define in the rule.
The operating system scan (OS-scan) type scans an operating system (and OS version) that an endpoint is running. This is a resource intensive scan.
Limitations of the NMAP tool that you use for an OS-scan may cause a scan of the operating system of endpoints to give unreliable results. For example, when scanning an operating system of network devices such as switches and routers, the NMAP OS-scan may provide an incorrect operating-system attribute for those devices. Cisco ISE displays the operating-system attribute, if the accuracy is not 100%.
For these devices, you can configure endpoint profiling policies that use the NMAP:operating-system attribute in their rules to have low certainty value conditions (Certainty Factor values). We recommend that whenever you create an endpoint profiling policy based on the NMAP:operating-system attribute, the profiling policy should contain an AND condition. For example, NMAP:operating-system contains 100.
Note Do not create an endpoint profiling policy based on the NMAP:operating-system attribute unless the accuracy is 100%.
The following NMAP command scans the operating system when you associate Scan OS with an endpoint profiling policy:
(Fast (limited port) scan). Specifies that you want to scan fewer ports than the default. Normally Nmap scans the most common 1000 ports for each scanned protocol. With -F, this is reduced to 100.
IP-address of an endpoint that is scanned.
SNMP Port Scan
The SNMPPortsAndOS-scan type scans an operating system (and OS version) that an endpoint is running and triggers an SNMP Query when SNMP ports (161 and 162) are open. It can be used for endpoints that are identified and matched initially with an Unknown profile for better classification.
The following NMAP command scans SNMP ports (UDP 161 and 162) when you associate the Scan SNMP Port with an endpoint profiling policy:
Table 22-9 NMAP Commands for an Endpoint Common Ports Scan
Both TCP connect scan and UDP scan.
-p <port ranges>
Scans TCP ports: 21,22,23,25,53,80,110,135,139,143, 443,445,3306,3389,8080 and UDP ports: 53,67,68,123,135,137, 138,139,161,445,500,520,631,1434,1900
IP address of an endpoint that is scanned.
The following table lists 30 common ports (15 TCP and 15 UDP ports) that NMAP uses for scanning.
Table 22-10 Common Ports
Cisco ISE Integration with Cisco NAC Appliance
Cisco ISE supports integration only with the Cisco Network Admission Control (NAC) Appliance Release 4.9 and is available when you have installed an Advanced or Wireless license in Cisco ISE.
The Cisco ISE profiler is similar to the Cisco Network Admission Control (NAC) Profiler that manages endpoints in a Cisco NAC deployment. This integration allows you to replace the existing Cisco NAC Profiler that is installed in a Cisco NAC deployment. It allows you to synchronize profile names from the Cisco ISE profiler and the result of endpoint classification in to the Cisco Clean Access Manager (CAM).
Cisco Clean Access Manager Configuration in Administration Nodes
Cisco ISE allows you to register multiple Clean Access Managers (CAMs) on a primary Administration node in a distributed deployment for REST APIs communication settings. The list of CAMs that is registered in Cisco ISE is the list to which all the profiler configuration changes are notified. The primary Administration node is responsible for all the communication between Cisco ISE and the Cisco NAC Appliance. You can configure CAMs only in the primary Administration node in Cisco ISE. The credentials that are used at the time of registering one or more CAMs in the primary Administration node are used to authenticate connectivity with CAMs.
The communication between Cisco ISE and the Cisco NAC Appliance is secure over Secure Sockets Layer (SSL). It is also bidirectional in nature, because Cisco ISE pushes the profiler configuration changes to CAMs, and CAMs periodically pull the list of MAC addresses of endpoints and their corresponding profiles and the list of all the profile names, from Cisco ISE.
You must export the contents of the X509 Certificate from the Clean Access Manager in Administration > Clean Access Manager > SSL, and import it into the primary Administration node under Administration > System > Certificates > Certificate Trust Store in Cisco ISE for a proper secure communication between Cisco ISE and CAM.
Cisco ISE Profiler and Cisco Clean Access Manager Communication
The Cisco ISE profiler notifies the profiler configuration changes to all the registered Clean Access Managers (CAMs) from the primary Administration node. It avoids duplicating notification in a Cisco ISE distributed deployment. It uses the REST APIs to notify the profiler configuration changes when endpoints are added or removed, and endpoint profiling policies changed, in the Cisco ISE database. During an import of endpoints, the Cisco ISE profiler notifies CAMs only after the import is complete.
The following REST API flow is implemented to push the profiler configuration changes to CAMs:
Cisco ISE profiler endpoint change push—When endpoints are profiled and there are changes in the profiles of endpoints in Cisco ISE, then the Cisco ISE profiler notifies all the registered CAMs about the changes in the endpoint profiles.
You can configure Cisco ISE in CAMs, which allows you to synchronize CAMs with Cisco ISE, depending on your Sync Settings in CAMs. You must create rules, where you can select one or more matching profiles from the list of Cisco ISE profiles and map endpoints to any one of the Access Types in CAMs. CAMs periodically retrieve endpoints and their corresponding profiles and the list of all the profile names, from the Cisco ISE profiler.
The following REST API flows are implemented to pull the profiler configuration changes from the Cisco ISE profiler:
NAC Manager endpoint pull—Pulls the list of MAC addresses of endpoints and their corresponding profiles of known endpoints.
NAC Manager profile pull—Pulls the profile names from the Cisco ISE profiler.
The Cisco ISE profiler notifies the Cisco ISE Monitoring persona of all the events that can be used to monitor and troubleshoot Cisco ISE and Cisco NAC Appliance Release 4.9 integration.
The Cisco ISE profiler log captures the following events for monitoring and troubleshooting integration:
Configuration changes for NAC Settings (Information)
NAC notification event failure (Error)
Adding Cisco Clean Access Managers
Integrating Cisco ISE with the Cisco NAC Appliance, Release 4.9 allows you to utilize the Cisco ISE profiling service in a Cisco NAC deployment. to utilize the Cisco ISE profiling service in a Cisco NAC deployment.
The NAC Managers page allows you to configure multiple Cisco Access Managers (CAMs), which provides an option to filter the CAMs that you have registered. This page lists the CAMs along with their names, descriptions, IP addresses, and the status that displays whether endpoint notification is enabled or not for those CAMs.
Creating Endpoints with Static Assignments of Policies and Identity Groups
You can create a new endpoint statically by using the MAC address of an endpoint in the Endpoints page. You can also choose an endpoint profiling policy and an identity group in the Endpoints page for static assignment.
You can import endpoints from a CSV file for which you have already exported endpoints from a Cisco ISE server, or a CSV file that you have created from Cisco ISE and updated with endpoint details.
The file format has to be in the format as specified in the default import template so that the list of endpoints appears as follows: MAC, Endpoint Policy, Endpoint Identity Group.
Both endpoint policy and endpoint identity group are optional for importing endpoints in a CSV file. If you want to import the endpoint identity group without the endpoint policy for endpoints, the values are still separated by the comma.
You can generate a template in which you can update endpoints that can be used to import endpoints. By default, you can use the Generate a Template link to create a CSV file in the Microsoft Office Excel application and save the file locally on your system. The file can be found in Administration > Identity Management > Identities > Endpoints > Import > Import From File.
You can use the Generate a Template link to create a template, and the Cisco ISE server will display the Opening template.csv dialog. This dialog allows you to open the default template.csv file, or save the template.csv file locally on your system. If you choose to open the template.csv file from the dialog, the file opens in the Microsoft Office Excel application.
The default template.csv file contains a header row that displays the MAC address, Endpoint Policy, and Endpoint Identity Group columns. You must update the MAC addresses of endpoints, endpoint profiling policies, and endpoint identity groups and save the file with a different file name that you can use to import endpoints.
Table 22-11 displays the header row in the template.csv file that is created when you use the Generate a Template link.
Table 22-11 CSV Template File
Endpoint Identity Group
Unknown Endpoints Reprofiled During Import
If the file used for import contains endpoints that have their MAC addresses, and their assigned endpoint profiling policies is the Unknown profile, then those endpoints are immediately reprofiled in Cisco ISE to the matching endpoint profiling policies during import. However, they are not statically assigned to the Unknown profile. If endpoints do not have endpoint profiling policies assigned to them in the CSV file, then they are assigned to the Unknown profile, and then reprofiled to the matching endpoint profiling policies.
Table 22-12 shows how Cisco ISE reprofiles Unknown profiles that match the Xerox_Device profile during import. It also shows how Cisco ISE reprofiles an endpoint that is unassigned.
Table 22-12 Unknown Profiles: Import from a File
Endpoint Profiling Policy Assigned Before Import in Cisco ISE
Endpoint Profiling Policy Assigned After Import in Cisco ISE
If no profile is assigned to an endpoint, then it is assigned to the Unknown profile, and also reprofiled to the matching profile.
Static Assignments of Policies and Identity Groups for Endpoints Retained During Import
If the file used for import contains endpoints that have their MAC addresses, and their assigned endpoint profiling policy is the static assignment, then they are not reprofiled during import.
Table 22-13 shows how Cisco ISE retains the Cisco-Device profile, the static assignment of an endpoint during import.
Table 22-13 Static Assignment: Import From a File
Endpoint Profiling Policy Assigned Before Import in Cisco ISE
Endpoint Profiling Policy Assigned After Import in Cisco ISE
Cisco-Device (static assignment)
Endpoints with Invalid Profiles Not Imported
If endpoints are assigned to invalid profiles in the file used for import, then they are not imported because there are no matching profiles in Cisco ISE.
Table 22-14 also shows that endpoints are not imported when they are assigned to invalid profiles in the CSV file.
Table 22-14 Invalid Profiles: Import from a File
Endpoint Profiling Policy Assigned Before Import in Cisco ISE
Endpoint Profiling Policy Assigned After Import in Cisco ISE
If an endpoint such as 00:00:00:00:01:05 is assigned to an invalid profile other than the profiles that are available in Cisco ISE, then Cisco ISE displays a warning message that the policy name is invalid and the endpoint will not be imported.
The endpoint is not imported because there is no matching profile in Cisco ISE.
You can import the MAC addresses, the associated profiles, and the endpoint identity groups of endpoints securely from an LDAP server.
Before You Begin
Before you begin to import endpoints, ensure that you have installed the LDAP server.
You have to configure the connection settings and query settings before you can import from an LDAP server. If the connection settings or query settings are configured incorrectly in Cisco ISE, then the “LDAP import failed:” error message appears.
Exporting Endpoints with Comma-Separated Values File
You can export selected or all endpoints from a Cisco ISE server to different Cisco ISE servers in a comma-separated values (CSV) file in which endpoints are listed with their MAC addresses, endpoint profiling policies, and endpoint identity groups to which they are assigned.
Export All is the default option. If endpoints are filtered in the Endpoints page, only those filtered endpoints are exported when you are using the Export All option. By default, the profiler_endpoints.csv is the CSV file and the Microsoft Office Excel is the default application to open the CSV file from the Opening profiler_endpoints.csv dialog box or to save the CSV file. For example, you can export selected endpoints or all endpoints in the profiler_endpoints.csv file, which you can use to import those endpoints.
Step 2 Click Export, and choose one of the following:
Export Selected —You can export only the selected endpoints in the Endpoints page.
Export All —By default, you can export all the endpoints in the Endpoints page.
Step 3 Click OK to save the profiler_endpoints.csv file.
Cisco ISE displays identified endpoints that connect to your network and use resources on your network in the Endpoints page. An endpoint is typically a network-capable device that connect to your network through wired and wireless network access devices and VPN. Endpoints can be personal computers, laptops, IP phones, smart phones, gaming consoles, printers, fax machines, and so on.
The MAC address of an endpoint, expressed in hexadecimal form, is always the unique representation of an endpoint, but you can also identify an endpoint with a varying set of attributes and the values associated to them, called an attribute-value pair. You can collect a varying set of attributes for endpoints based on the endpoint capability, the capability and configuration of the network access devices and the methods (probes) that you use to collect these attributes.
Dynamically Profiled Endpoints
When endpoints are discovered on your network, they can be profiled dynamically based on the configured profiling endpoint profiling policies, and assigned to the matching endpoint identity groups depending on their profiles.
Statically Profiled Endpoints
An endpoint can be profiled statically when you create an endpoint with its MAC address and associate a profile to it along with an endpoint identity group in Cisco ISE. Cisco ISE does not reassign the profiling policy and the identity group for statically assigned endpoints.
If you do not have a matching profiling policy for an endpoint, you can assign an unknown profiling policy (Unknown) and the endpoint therefore will be profiled as Unknown. The endpoint profiled to the Unknown endpoint policy requires that you create a profile with an attribute or a set of attributes collected for that endpoint. The endpoint that does not match any profile is grouped within the Unknown endpoint identity group.
Identified Endpoints Locally Stored in Policy Service Nodes Database
Cisco ISE writes identified endpoints locally in the Policy Service node database. After storing endpoints locally in the database, these endpoints are then made available (remote write) in the Administration node database only when significant attributes change in the endpoints, and replicated to the other Policy Service nodes database.
The following are the significant attributes:
When you change endpoint profile definitions in Cisco ISE, all endpoints have to be reprofiled. A Policy Service node that collects the attributes of endpoints is responsible for reprofiling of those endpoints.
When a Policy Service node starts collecting attributes about an endpoint for which attributes were initially collected by a different Policy Service node, then the endpoint ownership changes to the current Policy Service node. The new Policy Service node will retrieve the latest attributes from the previous Policy Service node and reconcile the collected attributes with those attributes that were already collected.
When a significant attribute changes in the endpoint, attributes of the endpoint are automatically saved in the Administration node database so that you have the latest significant change in the endpoint. If the Policy Service node that owns an endpoint is not available for some reasons, then the Administrator ISE node will reprofile an endpoint that lost the owner and you have to configure a new Policy Service node for such endpoints.
Policy Service Nodes in Cluster
Cisco ISE uses Policy Service node group as a cluster that allows to exchange endpoint attributes when two or more nodes in the cluster collect attributes for the same endpoint. We recommend to create clusters for all Policy Service nodes that reside behind a load balancer.
If a different node other than the current owner receives attributes for the same endpoint, it sends a message across the cluster requesting the latest attributes from the current owner to merge attributes and determine if a change of ownership is needed. If you have not defined a node group in Cisco ISE, it is assumed that all nodes are within one cluster.
There are no changes made to endpoint creation and replication in Cisco ISE. Only the change of ownership for endpoints is decided based on a list of attributes (white list) used for profiling that are built from static attributes and dynamic attributes. See Table 22-5.
Upon subsequent attributes collection, the endpoint is updated on the Administration node, if anyone of the following attributes changes:
When an endpoint is edited and saved in the Administration node, the attributes are retrieved from the current owner of the endpoint.
You can create, edit, or delete an endpoint identity group on the Endpoint Identity Groups page. You can only edit the name and description of the identity groups, and delete identity groups that you have created in the system. You cannot edit the name of the endpoint identity groups although their description are editable, and delete endpoint identity groups that are created by Cisco ISE in the system.
Identified Endpoints Grouped in Endpoint Identity Groups
Cisco ISE groups discovered endpoints into their corresponding endpoint identity groups based on the endpoint profiling policies. Profiling policies are hierarchical, and they are applied at the endpoint identify groups level in Cisco ISE. By grouping endpoints to endpoint identity groups, and applying profiling policies to endpoint identity groups, Cisco ISE enables you to determine the mapping of endpoints to the endpoint profiles by checking corresponding endpoint profiling policies.
Cisco ISE creates a set of endpoint identity groups by default, and allows you to create your own identity groups to which endpoints can be assigned dynamically or statically. You can create an endpoint identity group and associate the identity group to one of the system-created identity groups. You can also assign an endpoint that you create statically to any one of the identity groups that exists in the system, and the profiling service cannot reassign the identity group.
Default Endpoint Identity Groups Created for Endpoints
Cisco ISE creates the following five endpoint identity groups by default: Blacklist, GuestEndpoints, Profiled, RegisteredDevices, and Unknown. In addition, it creates two more identity groups, such as Cisco-IP-Phone and Workstation, which are associated to the Profiled (parent) identity group. A parent group is the default identity group that exists in the system.
Cisco ISE creates the following endpoint identity groups:
Blacklist—This endpoint identity group includes endpoints that are statically assigned to this group in Cisco ISE and endpoints that are blacklisted in the device registration portal. An authorization profile can be defined in Cisco ISE to permit, or deny network access to endpoints in this group.
GuestEndpoints—This endpoint identity group includes endpoints that are used by guest users.
Profiled—This endpoint identity group includes endpoints that match endpoint profiling policies except Cisco IP phones and workstations in Cisco ISE.
RegisteredDevices—This endpoint identity group includes endpoints, which are registered devices that are added by an employee through the devices registration portal. The profiling service continues to profile these devices normally when they are assigned to this group. Endpoints are statically assigned to this group in Cisco ISE, and the profiling service cannot reassign them to any other identity group. These devices will appear like any other endpoint in the endpoints list. You can edit, delete, and blacklist these devices that you added through the device registration portal from the endpoints list in the Endpoints page in Cisco ISE. Devices that you have blacklisted in the device registration portal are assigned to the Blacklist endpoint identity group, and an authorization profile that exists in Cisco ISE redirects blacklisted devices to an URL, which displays “Unauthorised Network Access”, a default portal page to the blacklisted devices.
Unknown—This endpoint identity group includes endpoints that do not match any profile in Cisco ISE.
In addition to the above system created endpoint identity groups, Cisco ISE creates the following endpoint identity groups, which are associated to the Profiled identity group:
Cisco-IP-Phone—An identity group that contains all the profiled Cisco IP phones on your network.
Workstation—An identity group that contains all the profiled workstations on your network.
Endpoint Identity Groups Created for Matched Endpoint Profiling Policies
If you have an endpoint policy that matches an existing policy, then the profiling service can create a matching endpoint identity group. This identity group becomes the child of the Profiled endpoint identity group. When you create an endpoint policy, you can check the Create Matching Identity Group check box in the Profiling Policies page to create a matching endpoint identity group. You cannot delete the matching identity group unless the mapping of the profile is removed.
Adding Static Endpoints in Endpoint Identity Groups
You can add or remove statically added endpoints in any endpoint identity group.
You can add endpoints from the Endpoints widget only to a specific identity group. If you add an endpoint to the specific endpoint identity group, then the endpoint is moved from the endpoint identity group where it was dynamically grouped earlier.
Upon removal from the endpoint identity group where you recently added an endpoint, the endpoint is reprofiled back to the appropriate identity group. You do not delete endpoints from the system but only remove them from the endpoint identity group.
Dynamic Endpoints Reprofiled After Adding or Removing in Identity Groups
If an endpoint identity group assignment is not static, then endpoints are reprofiled after you add or remove them from an endpoint identity group. Endpoints that are identified dynamically by the ISE profiler appear in appropriate endpoint identity groups. If you remove dynamically added endpoints from an endpoint identity group, Cisco ISE displays a message that you have successfully removed endpoints from the identity group but reprofiles them back in the endpoint identity group.
Endpoint Identity Groups Used in Authorization Rules
You can effectively use endpoint identity groups in the authorization policies to provide appropriate network access privileges to the discovered endpoints. For example, an authorization rule for all types of Cisco IP Phones is available by default in Cisco ISE in the following location: Policy > Authorization > Standard.
You must ensure that the endpoint profiling policies are either standalone policies (not a parent to other endpoint profiling policies), or their parent policies of the endpoint profiling policies are not disabled.
Profiler Feed Service
With the introduction of the profiler feed service, the profiler conditions, exception actions, and NMAP scan actions are classified as Cisco-provided or administrator-created (see the System Type attribute) in Cisco ISE. Also, the endpoint profiling policies are classified as Cisco provided, administrator created, or administrator modified (see the System Type attribute).
You can perform different operations on the profiler conditions, exception actions, NMAP scan actions, and endpoint profiling policies depending on the System Type attribute. You cannot edit or delete Cisco-provided conditions, exception actions, and nmap scan actions. Endpoint policies that are provided by Cisco cannot be deleted. When policies are edited, they are considered as administrator-modified. when administrator-modified policies are deleted, they are replaced by the up-to-date version of the Cisco-provided policy that it was based on.
You can retrieve new and updated endpoint profiling policies and the updated OUI database as a feed from a designated Cisco feed server through a subscription into Cisco ISE. You can also receive e-mail notifications to the e-mail address as an administrator of Cisco ISE that you have configured for applied, success, and failure messages. You can also provide additional subscriber information to receive notifications. You can send the subscriber information back to Cisco for maintaining the records and they are treated as privileged and confidential.
By default, the profiler feed service is disabled, and it requires an Advanced license to enable the service. When you enable the profiler feed service, Cisco ISE downloads the feed service policies and OUI database updates every day at 1:00 A.M of the local Cisco ISE server time zone. Cisco ISE automatically applies these downloaded feed server policies, which also stores the set of changes so that you can revert these changes back to the previous state. When you revert from the set of changes that you last applied, endpoint profiling policies that are newly added are removed and endpoint profiling policies that are updated are reverted to the previous state. In addition, the profiler feed service is automatically disabled.
When the updates occur, only the Cisco Provided profiling policies are updated and the endpoint profiling policies which were modified by the previous update. Cisco Provided disabled profiling policies are also updated but they remain disabled. Administrator Created or Administrator Modified profiling policies are not overwritten. If you want to revert any Administrator Modified endpoint profiling policy to any Cisco Provided endpoint profiling policy, then you must delete the Administrator Modified endpoint profiling policy or revert the Administrator Modified endpoint profiling policy to the previous Cisco Provided endpoint profiling policy.
The designated Cisco feed server downloads the updated OUI database from the following location: http://standards.ieee.org/develop/regauth/oui/oui.txt, which is the list of vendors associated to the MAC OUI. The updated OUI database is available for any ISE deployment as a feed that Cisco ISE downloads to its own database. Cisco ISE updates endpoints and then starts reprofiling endpoints.
The designated Cisco feed server is located at https://ise.cisco.com:8443/feedserver/. If you have any issues accessing the service, ensure that your network security components (like a firewall or proxy server, for example) allow direct access to this URL.
Configuring Profiler Feed Service
You can configure Cisco ISE to receive updates for new, updated endpoint profiling policies, and MAC OUI as a feed from a designated Cisco feed server. Enabling the profiler feed service allows Cisco ISE to check for new and updated profiles that are created since the last feed service update, and MAC OUI. If feed service is unavailable or other errors have occured, then they are reported in the Operations Audit report.
You can configure Cisco ISE to send the feed service usage report back to Cisco as an administrator of Cisco ISE and provide an alternate Cisco ISE administrator details (optional attributes) in the Feed Service Subscriber Information section.
The CoA type in a profiling policy downloaded from the feed server is locally configured and not downloaded. If you change the CoA type in the downloaded policy, the state will not be updated. It remains as a system policy and you need to update the policy via the feed server.
Before You Begin
You can configure the profiler feed service only from the Cisco ISE Admin portal in a distributed deployment or in a standalone ISE node.
Cisco ISE requires the feed server certificate in order to communicate as the feed service client interacts with the feed service platform through HTTPS. You can download the certificate from the Certificate Store by navigating to Administration > System > Certificates.
You must set up a Simple Mail Transfer Protocol (SMTP) server to send e-mail notifications from the Admin portal (Administration > System > Settings).
Step 2 Check the Enable Profiler Feed Service check box.
Step 3 Enter time in HH:MM format (local time zone of the Cisco ISE server) in the Feed Service Scheduler section. By default, Cisco ISE profiler feed service is scheduled at 1.00 AM every day.
Step 4 Check the Notify administrator when download occurs check box in the Administrator Notification Options section and enter your e-mail address as an administrator of Cisco ISE in the Administrator email address text box.
Step 5 Check the Provide subscriber information to Cisco check box in the Feed Service Subscriber Information section to send the feed server usage report back to Cisco.
Instructs Cisco ISE to contact Cisco feed server for new and updated profiles created since the last feed service update. This reprofiles all endpoints in the system, which may cause an increase in the system load. Due to updated endpoint profiling policies, there may be changes in the authorization policy for some endpoints that are currently connected to Cisco ISE.
The Update Now button is disabled when you update new and updated profiles created since the last feed service and enabled only after the download is completed. You must navigate away from the Profiler Feed Service Configuration page and return to this page.
You can revert endpoint profiling policies that were updated in the previous update and remove endpoint profiling policies that are newly added through the previous update of the profiler feed service but OUI updates are not changed.
An endpoint profiling policy, if modified after an update from the feed server is not changed in the system.
Cisco ISE provides you with various reports on endpoint profiling, and troubleshooting tools that you can use to manage your network. You can generate reports for historical as well as current data. You may be able to drill down on a part of the report to view more details. For large reports, you can also schedule reports and download them in various formats.
You can run the following reports for endpoints from Operations > Reports > Endpoints and Users: