Port, Protocol, and ICMP Code Rule Conditions

Port conditions match traffic based on the source and destination ports. Depending on the rule type, “port” can represent any of the following:

  • TCP and UDP—You can control TCP and UDP traffic based on the port. The system represents this configuration using the protocol number in parentheses, plus an optional associated port or port range. For example: TCP(6)/22.

  • ICMP—You can control ICMP and ICMPv6 (IPv6-ICMP) traffic based on its internet layer protocol plus an optional type and code. For example: ICMP(1):3:3.

  • Protocol—You can control traffic using other protocols that do not use ports.

Leave matching criteria empty whenever possible, especially those for security zones, network objects, and port objects. When you specify multiple criteria, the system must match against every combination of the contents of the criteria you specify.

Best Practices for Port-Based Rules

Specifying ports is the traditional way to target applications. However, applications can be configured to use unique ports to bypass access control blocks. Thus, whenever possible, use application filtering criteria rather than port criteria to target traffic. Note that application filtering is not available in prefilter rules.

Application filtering is also recommended for applications, like FTP, that open separate channels dynamically for control vs. data flow. Using port-based access control rules can prevent these kinds of applications from performing correctly, and could result in blocking desirable connections.

Using Source and Destination Port Constraints

If you add both source and destination port constraints, you can only add ports that share a single transport protocol (TCP or UDP). For example, if you add DNS over TCP as a source port, you can add Yahoo Messenger Voice Chat (TCP) as a destination port but not Yahoo Messenger Voice Chat (UDP).

If you add only source ports or only destination ports, you can add ports that use different transport protocols. For example, you can add both DNS over TCP and DNS over UDP as destination port conditions in a single access control rule.

Matching Non-TCP Traffic with Port Conditions

You can match non-port-based protocols. By default, if you do not specify a port condition, you are matching IP traffic. Although you can configure port conditions to match non-TCP traffic, there are some restrictions:

  • Access control rules—For Classic devices, you can match GRE-encapsulated traffic with an access control rule by using the GRE (47) protocol as a destination port condition. To a GRE-constrained rule, you can add only network-based conditions: zone, IP address, port, and VLAN tag. Also, the system uses outer headers to match all traffic in access control policies with GRE-constrained rules. For threat defense devices, use tunnel rules in the prefilter policy to control GRE-encapsulated traffic.

  • Decryption rules—These rules support TCP port conditions only.

  • IMCP echo—A destination ICMP port with the type set to 0 or a destination ICMPv6 port with the type set to 129 only matches unsolicited echo replies. ICMP echo replies sent in response to ICMP echo requests are ignored. For a rule to match on any ICMP echo, use ICMP type 8 or ICMPv6 type 128.