This document describes the Firewall Threat Defense (FTD) Modular Policy Framework (MPF)
There are not specific requirements for this document.
The information in this document is based on these software and hardware versions:
The information in this document was created from the devices in a specific lab environment. All of the devices used in this document started with a cleared (default) configuration. If your network is live, ensure that you understand the potential impact of any command.
FTD Data Plane Overview
FTD is a unified software image that consists of 2 main engines:

The LINA Datapath and the Snort Engine are the main parts of the FTD’s Data Plane.
MPF uses these components:
Regarding feature directionality, consult the ASA configuration guide:
The features related to the FTD are highlighted:


The default MPF configuration (10.0.0):
firewall# show run policy-map ! !
policy-map type inspect dns preset_dns_map
parameters
message-length maximum client auto
message-length maximum 512
no tcp-inspection
policy-map type inspect ip-options UM_STATIC_IP_OPTIONS_MAP
parameters
eool action allow
nop action allow
router-alert action allow
policy-map global_policy
class inspection_default
inspect dns preset_dns_map
inspect ftp
inspect h323 h225
inspect h323 ras
inspect rsh
inspect rtsp
inspect sqlnet
inspect skinny
inspect sunrpc
inspect sip
inspect netbios
inspect tftp
inspect icmp
inspect icmp error
inspect ip-options UM_STATIC_IP_OPTIONS_MAP
class class_snmp
inspect snmp
class class-default
set connection advanced-options UM_STATIC_TCP_MAP
firewall# show run class-map !
class-map inspection_default
match default-inspection-traffic
class-map class_snmp
match port udp eq 4161
! firewall# show run service-policy service-policy global_policy global
The requirement in this task is to disable SIP inspection in the FTD LINA engine. One reason can be a policy requirement or a software defect related to SIP that affects transit traffic.
Solution
Before disabling SIP inspection first confirm that it s applied to transit traffic:
firewall# packet-tracer input INSIDE udp 172.16.1.1 5060 172.16.3.1 5060
...
Phase: 8
Type: INSPECT
Subtype: inspect-sip
Result: ALLOW
Elapsed time: 34788 ns
Config:
class-map inspection_default
match default-inspection-traffic
policy-map global_policy
class inspection_default
inspect sip
service-policy global_policy global
Additional Information:
...
Result:
input-interface: INSIDE(vrfid:0)
input-status: up
input-line-status: up
output-interface: OUTSIDE1(vrfid:0)
output-status: up
output-line-status: up
Action: allow
Time Taken: 326018 ns
There are 2 ways to globally disable SIP inspection:
Solution 1: Disable SIP from FTD CLISH CLI
> configure inspection sip disable Building configuration... Cryptochecksum: ef7528dc 7338986d 6714a3a2 4770528e 7818 bytes copied in 0.250 secs [OK]
Verification
> show running-config policy-map | include sip >
Solution 2: Disable SIP using FlexConfig
On FMC navigate to Devices > FlexConfig and create a FlexConfig object:

policy-map global_policy class inspection_default no inspect sip
Apply the FlexConfig policy and select Preview Config in order to preview it:

Finally, Deploy the policy.
Verification
firewall# show run policy-map | include sip firewall#
Note – You need to clear the existing SIP connection from LINA connection table so that the connections are re-established without SIP inspection. You can use this command in order to verify the existing SIP connections:
firewall# show conn port 5060
In this task the requirement is to disable SIP inspection for traffic between these networks:
One reason to do this can be a software defect related to SIP that affects transit traffic
Solution
Use FlexConfig.
Step 1
Navigate to Objects > Access List > Extended and create an extended access list that matches the interesting traffic. You have to use the Block action since the goal to exclude the specific traffic. Additionally, add an Allow rule to match the rest of the traffic:

Step 2
Create a FlexConfig object with a class-map that matches the SIP Access Control List (ACL) and apply it into the global_policy:

The configured FlexConfig object:
class-map SIP_CMAP
match access-list $SIP_flows
policy-map global_policy
class inspection_default
no inspect sip
class SIP_CMAP
inspect sip
Note
When configuring the permit ACL try to be as specific as possible (for example, put protocol ports) to avoid any potential CPU impact. The example in this task does not specify protocol ports and can be avoided in production.
Verification 1
firewall# show run policy-map | begin global
policy-map global_policy
class inspection_default
inspect dns preset_dns_map
inspect ftp
inspect h323 h225
inspect h323 ras
inspect rsh
inspect rtsp
inspect sqlnet
inspect skinny
inspect sunrpc
inspect netbios
inspect tftp
inspect icmp
inspect icmp error
inspect ip-options UM_STATIC_IP_OPTIONS_MAP
class class_snmp
inspect snmp
class SIP_CMAP
inspect sip
class class-default
set connection advanced-options UM_STATIC_TCP_MAP
firewall# show run class-map
!
class-map SIP_CMAP
match access-list SIP_flows
class-map inspection_default
match default-inspection-traffic
class-map class_snmp
match port udp eq 4161
firewall# show run access-list SIP_flows
access-list SIP_flows extended deny ip 172.16.1.0 255.255.255.0 172.16.3.0 255.255.255.0
access-list SIP_flows extended permit ip any any
Verification 2
Traffic that is not inspected by SIP inspection has deny=true:
firewall# packet-tracer input INSIDE udp 172.16.1.1 5060 172.16.3.1 5060 detail | begin INSPECT
Type: INSPECT
Subtype: inspect-sip
Result: ALLOW
Elapsed time: 37910 ns
Config:
class-map SIP_CMAP
match access-list SIP_flows
policy-map global_policy
class SIP_CMAP
inspect sip
service-policy global_policy global
Additional Information:
Forward Flow based lookup yields rule:
in id=0x14af42cfa810, priority=70, domain=inspect-sip, deny=true
hits=1, user_data=0x000014af4570bea0, cs_id=0x0, use_real_addr, flags=0x0, protocol=0
src ip/id=172.16.1.0, mask=255.255.255.0, port=0, tag=any
dst ip/id=172.16.3.0, mask=255.255.255.0, port=0, tag=any,
dscp=0x0, input_ifc=INSIDE(vrfid:0), output_ifc=any
...
Traffic that is inspected by SIP inspection has deny=false:
firewall# packet-tracer input INSIDE udp 172.16.2.1 5060 172.16.3.1 5060 detail | begin INSPECT
Type: INSPECT
Subtype: inspect-sip
Result: ALLOW
Elapsed time: 34788 ns
Config:
class-map SIP_CMAP
match access-list SIP_flows
policy-map global_policy
class SIP_CMAP
inspect sip
service-policy global_policy global
Additional Information:
Forward Flow based lookup yields rule:
in id=0x14af459099d0, priority=70, domain=inspect-sip, deny=false
hits=1, user_data=0x000014af4570bea0, cs_id=0x0, use_real_addr, flags=0x0, protocol=0
src ip/id=0.0.0.0, mask=0.0.0.0, port=0, tag=any
dst ip/id=0.0.0.0, mask=0.0.0.0, port=0, tag=any,
...
Verification 3
The "sip" inspect counter increases when a packet is inspected by the firewall:
firewall# show service-policy inspect sip
Global policy:
Service-policy: global_policy
Class-map: inspection_default
Class-map: class_snmp
Class-map: SIP_CMAP
Inspect: sip , packet 2, lock fail 0, drop 0, reset-drop 0, 5-min-pkt-rate 0 pkts/sec, v6-fail-close 0 sctp-drop-override 0
tcp-proxy: bytes in buffer 0, bytes dropped 0
...
firewall# packet-tracer input INSIDE udp 172.16.2.1 5060 172.16.3.1 5060
firewall#show service-policy inspect sip
Global policy:
Service-policy: global_policy
Class-map: inspection_default
Class-map: class_snmp
Class-map: SIP_CMAP
Inspect: sip , packet 3, lock fail 0, drop 0, reset-drop 0, 5-min-pkt-rate 0 pkts/sec, v6-fail-close 0 sctp-drop-override 0
tcp-proxy: bytes in buffer 0, bytes dropped 0
...
In this task the requirement is to enable TCP state bypass for traffic between these networks:
In general, it is not recommended to use TCP state bypass, but it can be used as a temporary workaround for handling asymmetric flows.
Solution 1
Step 1
Create an extended ACL that matches the interesting traffic:

Step 2
Edit the Access Control Policy (ACP) assigned to the FTD, select the Advanced Settings tab and edit the Threat Defense Service Policy. Select Add Rule and Next.
Step 3
Select the extended ACL:

Step 4

Step 5
Select Finish, OK, Save and Deploy.
The result:
firewall# show run policy-map global_policy
!
policy-map global_policy
class inspection_default
inspect dns preset_dns_map
inspect ftp
inspect h323 h225
inspect h323 ras
inspect rsh
inspect rtsp
inspect sqlnet
inspect skinny
inspect sunrpc
inspect netbios
inspect tftp
inspect icmp
inspect icmp error
inspect ip-options UM_STATIC_IP_OPTIONS_MAP
class class_map_TCP_Bypass
set connection random-sequence-number disable
set connection advanced-options tcp-state-bypass
class class_snmp
inspect snmp
class class-default
set connection advanced-options UM_STATIC_TCP_MAP
Note: In previous FMC releases like 6.x you can use FlexConfig to configure TCP state bypass. In newer versions this is not supported:

Verification
firewall# packet-tracer input INSIDE tcp 172.16.2.1 1111 172.16.3.1 80 detail | begin CONN
Type: CONN-SETTINGS
Subtype:
Result: ALLOW
Elapsed time: 334 ns
Config:
class-map class_map_TCP_Bypass
match access-list TCP_Bypass
policy-map global_policy
class class_map_TCP_Bypass
set connection conn-max 0 embryonic-conn-max 0 random-sequence-number disable syn-cookie-mss 1380
set connection advanced-options tcp-state-bypass
service-policy global_policy global
Additional Information:
Forward Flow based lookup yields rule:
in id=0x14af45906b70, priority=7, domain=conn-set, deny=false
hits=1, user_data=0x000014af45906df0, cs_id=0x0, use_real_addr, flags=0x0, protocol=0
src ip/id=172.16.2.0, mask=255.255.255.0, port=0, tag=any
dst ip/id=172.16.3.0, mask=255.255.255.0, port=0, tag=any,
dscp=0x0, input_ifc=INSIDE(vrfid:0), output_ifc=any
...
Prerequisite
Configure static NAT on FTD so that IP 172.16.1.1 located behind INSIDE interface appears as 198.51.100.200 on OUTSIDE1 hosts:

Then, run a traceroute from ISP1 to 198.51.100.200 (host 172.16.1.1):

router1# traceroute vrf VRF-201 198.51.100.200
Type escape sequence to abort.
Tracing the route to 198.51.100.200
VRF info: (vrf in name/id, vrf out name/id)
1 192.0.2.99 1 msec 1 msec *
Requirement
Modify the FTD configuration so that the traceroute matches this output:
router1# traceroute vrf VRF-201 198.51.100.200
Type escape sequence to abort.
Tracing the route to 198.51.100.200
VRF info: (vrf in name/id, vrf out name/id)
1 198.51.100.129 1 msec 1 msec *
2 198.51.100.200 1 msec 2 msec *
Solution
The solution includes two configuration steps:
1. Decrement the TTL:

After this change, the traceroute reveals the firewall hop:
router1# traceroute vrf VRF-201 198.51.100.200
Type escape sequence to abort.
Tracing the route to 198.51.100.200
VRF info: (vrf in name/id, vrf out name/id)
1 198.51.100.129 1 msec 1 msec *
2 192.0.2.99 1 msec 1 msec *
2. Disable ICMP Error inspection:

policy-map global_policy class inspection_default no inspect icmp error
Verification
The traceroute shows the translated NAT IP address of the remote host and the FTD interface IP address:
router1# traceroute vrf VRF-201 198.51.100.200
Type escape sequence to abort.
Tracing the route to 198.51.100.200
VRF info: (vrf in name/id, vrf out name/id)
1 198.51.100.129 1 msec 1 msec *
2 198.51.100.200 1 msec 2 msec *
Requirement
Change the timeout to 1 week for this flow:
Solution
To set timeout per flow you need to use Service Policy.
Step 1
Navigate to Objects > Access List and create an extended ACL that matches the interesting traffic:

Step 2
Configure an MPF policy that uses the ACL that was created in step 1:

Set the connection idle timeout:

Remove the rule from the previous task since it overlaps with the new requirement:

Verification
The deployed policy-map configuration:
policy-map global_policy
class inspection_default
inspect dns preset_dns_map
inspect ftp
inspect h323 h225
inspect h323 ras
inspect rsh
inspect rtsp
inspect sqlnet
inspect skinny
inspect sunrpc
inspect netbios
inspect tftp
inspect icmp
inspect ip-options UM_STATIC_IP_OPTIONS_MAP
inspect sip
class class_map_TCP_conn_timeout_ACL
set connection timeout idle 168:00:00
class class_snmp
inspect snmp
class class-default
set connection advanced-options UM_STATIC_TCP_MAP
Initiate a new TCP connection from 172.16.1.1 to 172.16.5.1 and check the connection table of the FTD:
firewall# show conn long address 172.16.5.1
...
TCP OUTSIDE2: 172.16.5.1/23 (172.16.5.1/23) INSIDE: 172.16.1.1/29389 (172.16.1.1/29389), flags UIoN1N7, idle 5s, uptime 7s, timeout 7D0h, bytes 349, flow id 72, Snort id 6, rule id 268439559, Rx-RingNum 27, Internal-Data0/1
Initiator: 172.16.1.1, Responder: 172.16.5.1
Connection lookup keyid: 890
Prerequisite
Configure a BGP session through the FTD. The BGP session nees to use authentication.

Verification
With the default FTD configuration, the BGP session is not established. On the router you can see:
router1#
*May 21 07:51:23.595: %TCP-6-BADAUTH: Invalid MD5 digest from 192.0.2.99(24591) to 198.51.100.250(179) tableid - 3
*May 21 07:51:25.595: %TCP-6-BADAUTH: Invalid MD5 digest from 192.0.2.99(24591) to 198.51.100.250(179) tableid - 3
*May 21 07:51:29.595: %TCP-6-BADAUTH: Invalid MD5 digest from 192.0.2.99(24591) to 198.51.100.250(179) tableid - 3
On the FTD you see that both sides fail to establish the BGP TCP connection (the connection flags denote that only TCP SYN packets are being received):
firewall# show conn port 179
3 in use, 16 most used
Inspect Snort:
preserve-connection: 2 enabled, 0 in effect, 15 most enabled, 0 most in effect
TCP OUTSIDE1 198.51.100.250:41090 INSIDE 192.0.2.99:179, idle 0:00:00, bytes 0, flags aA N1
TCP OUTSIDE1 198.51.100.250:179 INSIDE 192.0.2.99:53629, idle 0:00:02, bytes 0, flags aA N1
Solution
To allow an authenticated BGP session through the FTD these 2 conditions must to be met:
TCP MD5 option is allowed by default:

firewall# show run all tcp-map
!
tcp-map UM_STATIC_TCP_MAP
no check-retransmission
no checksum-verification
exceed-mss allow
queue-limit 0 timeout 4
reserved-bits allow
syn-data allow
synack-data drop
invalid-ack drop
seq-past-window drop
tcp-options range 6 7 allow
tcp-options range 9 18 allow
tcp-options range 20 255 allow
tcp-options selective-ack allow
tcp-options timestamp allow
tcp-options window-scale allow
tcp-options mss allow
tcp-options md5 allow
ttl-evasion-protection
urgent-flag allow
window-variation allow-connection
Globally disable TCP Initial Sequence Number (ISN) randomization:
> configure tcp-randomization disable Building configuration... Cryptochecksum: f8ac5587 7ccc635e bff886a1 bcab820c 8284 bytes copied in 0.260 secs [OK] >
or (the preferred method) you create an extended access list that matches the BGP connection:

and disable the TCP sequence number randomization using the Threat Defense Service Policy:

Verification
The deployed policy-map configuration:
policy-map global_policy
class inspection_default
inspect dns preset_dns_map
inspect ftp
inspect h323 h225
inspect h323 ras
inspect rsh
inspect rtsp
inspect sqlnet
inspect skinny
inspect sunrpc
inspect netbios
inspect tftp
inspect icmp
inspect ip-options UM_STATIC_IP_OPTIONS_MAP
inspect sip
class class_map_BGP_ACL
set connection random-sequence-number disable
class class_snmp
inspect snmp
class class-default
set connection advanced-options UM_STATIC_TCP_MAP
The BGP session is established through FTD:
firewall# show conn long port 179
...
TCP OUTSIDE1: 198.51.100.250/49863 (198.51.100.250/49863) INSIDE: 192.0.2.99/179 (192.0.2.99/179), flags UIO N1N7, idle 44s, uptime 1m40s, timeout 1h0m, bytes 274, flow id 111, Snort id 3, rule id 268439559, Rx-RingNum 29, Internal-Data0/1
Initiator: 198.51.100.250, Responder: 192.0.2.99
Connection lookup keyid: 83487134
Requirement
Configure DCD on FTD for TCP traffic destined to host 172.16.3.1.
Solution
DCD is documented at:
1. Navigate to Objects > Access-List and create an access-list that matches the interesting traffic.
2. Edit the ACP assigned to your firewall, navigate to Advanced options and select Threat Defense Service Policy to enable DCD:

The deployed configuration:
access-list DCD_ACL extended permit object-group ProxySG_ExtendedACL_81604390279 any host 172.16.3.1
!
class-map class_map_DCD_ACL
match access-list DCD_ACL
policy-map global_policy
class class_map_DCD_ACL
set connection timeout dcd
How it works
Configure FTD captures to see the backend operation:
firewall# capture CAPI interface INSIDE match tcp host 172.16.3.1 any
firewall# capture CAPO interface OUTSIDE1 match tcp host 172.16.3.1 any
Establish a TCP connection through the firewall:
firewall# show conn long address 172.16.3.1 | begin 172.16.3.1
TCP OUTSIDE1: 172.16.3.1/23 (172.16.3.1/23) INSIDE: 192.0.2.99/23241 (192.0.2.99/23241), flags UIO N1N7, idle 1m18s, uptime 1m22s, timeout 5m0s, bytes 129, flow id 127, Snort id 4, rule id 268439559, Rx-RingNum 13, Internal-Data0/1
Initiator: 192.0.2.99, Responder: 172.16.3.1
DCD probes sent: Initiator 0, Responder 0 Connection lookup keyid: 76292550
Initially, there are no DCD packets shown in the firewall captures:
firewall# show capture
capture CAPI type raw-data interface INSIDE [Capturing - 0 bytes]
match tcp host 172.16.3.1 any
capture CAPO type raw-data interface OUTSIDE1 [Capturing - 0 bytes]
match tcp host 172.16.3.1 any
When an idle connection reaches the idle timeout, the FTD sends spoofed TCP ACK messages to the source and destination:
firewall# show conn long address 172.16.3.1 | begin 172.16.3.1
TCP OUTSIDE1: 172.16.3.1/23 (172.16.3.1/23) INSIDE: 192.0.2.99/23241 (192.0.2.99/23241), flags UIO N1N7, idle 4m59s, uptime 5m3s, timeout 5m0s, bytes 129, flow id 127, Snort id 4, rule id 268439559, Rx-RingNum 13, Internal-Data0/1
Initiator: 192.0.2.99, Responder: 172.16.3.1
DCD probes sent: Initiator 0, Responder 0 Connection lookup keyid: 76292550
firewall# show conn long address 172.16.3.1 | begin 172.16.3.1
TCP OUTSIDE1: 172.16.3.1/23 (172.16.3.1/23) INSIDE: 192.0.2.99/23241 (192.0.2.99/23241), flags UIO N1N7, idle 0s, uptime 5m3s, timeout 15s, bytes 129, flow id 127, Snort id 4, rule id 268439559, Rx-RingNum 13, Internal-Data0/1
Initiator: 192.0.2.99, Responder: 172.16.3.1
DCD probes sent: Initiator 1, Responder 0 Connection lookup keyid: 76292550
firewall# show conn long address 172.16.3.1 | begin 172.16.3.1
TCP OUTSIDE1: 172.16.3.1/23 (172.16.3.1/23) INSIDE: 192.0.2.99/23241 (192.0.2.99/23241), flags UIO N1N7, idle 0s, uptime 5m4s, timeout 5m0s, bytes 129, flow id 127, Snort id 4, rule id 268439559, Rx-RingNum 13, Internal-Data0/1
Initiator: 192.0.2.99, Responder: 172.16.3.1
DCD probes sent: Initiator 1, Responder 1 Connection lookup keyid: 76292550
If both reply, it resets the idle timer:
firewall# show capture CAPI
3 packets captured
1: 09:01:30.433952 802.1Q vlan#101 P0 172.16.3.1.23 > 192.0.2.99.23241: . ack 3271882019 win 32757
2: 09:01:30.434334 802.1Q vlan#101 P0 192.0.2.99.23241 > 172.16.3.1.23: . ack 1746306341 win 32746
3: 09:01:30.955654 802.1Q vlan#101 P0 172.16.3.1.23 > 192.0.2.99.23241: . ack 3271882019 win 32757
3 packets shown
firewall# show capture CAPO
3 packets captured
1: 09:01:30.434364 802.1Q vlan#201 P0 192.0.2.99.23241 > 172.16.3.1.23: . ack 111661490 win 32746
2: 09:01:30.955288 802.1Q vlan#201 P0 192.0.2.99.23241 > 172.16.3.1.23: . ack 111661490 win 32746
3: 09:01:30.955639 802.1Q vlan#201 P0 172.16.3.1.23 > 192.0.2.99.23241: . ack 3875469573 win 32757
3 packets shown
firewall# show conn long address 172.16.3.1 | begin 172.16.3.1
TCP OUTSIDE1: 172.16.3.1/23 (172.16.3.1/23) INSIDE: 192.0.2.99/23241 (192.0.2.99/23241), flags UIO N1N7, idle 1m29s, uptime 6m33s, timeout 5m0s, bytes 129, flow id 127, Snort id 4, rule id 268439559, Rx-RingNum 13, Internal-Data0/1
Initiator: 192.0.2.99, Responder: 172.16.3.1
DCD probes sent: Initiator 1, Responder 1 Connection lookup keyid: 76292550
Note: DCD does not work on offloaded connections ('o' flag).
| Revision | Publish Date | Comments |
|---|---|---|
1.0 |
22-May-2026
|
Initial Release |