About Advanced Interface Configuration
This section describes advanced interface settings.
About MAC Addresses
You can manually assign MAC addresses to override the default. For multiple context mode, you can automatically generate unique MAC addresses (for all interfaces assigned to a context) and single context mode (for subinterfaces)..
Note |
You might want to assign unique MAC addresses to subinterfaces defined on the ASA, because they use the same burned-in MAC address of the parent interface. For example, your service provider might perform access control based on the MAC address. Also, because IPv6 link-local addresses are generated based on the MAC address, assigning unique MAC addresses to subinterfaces allows for unique IPv6 link-local addresses, which can avoid traffic disruption in certain instances on the ASA device. |
Default MAC Addresses
Default MAC address assignments depend on the type of interface.
-
Physical interfaces—The physical interface uses the burned-in MAC address.
-
VLAN interfaces (Firepower 1010 and Secure Firewall 1210/1220)—Routed firewall mode: All VLAN interfaces share a MAC address. Ensure that any connected switches can support this scenario. If the connected switches require unique MAC addresses, you can manually assign MAC addresses. See Configure the Manual MAC Address, MTU, and TCP MSS.
Transparent firewall mode: Each VLAN interface has a unique MAC address. You can override the generated MAC addresses if desired by manually assigning MAC addresses. See Configure the Manual MAC Address, MTU, and TCP MSS.
-
EtherChannels (Firepower Models)—For an EtherChannel, all interfaces that are part of the channel group share the same MAC address. This feature makes the EtherChannel transparent to network applications and users, because they only see the one logical connection; they have no knowledge of the individual links. The port-channel interface uses a unique MAC address from a pool; interface membership does not affect the MAC address.
-
EtherChannels (ASA Models)—The port-channel interface uses the lowest-numbered channel group interface MAC address as the port-channel MAC address. Alternatively you can configure a MAC address for the port-channel interface. We recommend configuring a unique MAC address in case the group channel interface membership changes. If you remove the interface that was providing the port-channel MAC address, then the port-channel MAC address changes to the next lowest numbered interface, thus causing traffic disruption.
-
Subinterfaces—All subinterfaces of a physical interface use the same burned-in MAC address. You might want to assign unique MAC addresses to subinterfaces. For example, your service provider might perform access control based on the MAC address. Also, because IPv6 link-local addresses are generated based on the MAC address, assigning unique MAC addresses to subinterfaces allows for unique IPv6 link-local addresses, which can avoid traffic disruption in certain instances on the ASA.
Automatic MAC Addresses
In multiple context mode, auto-generation assigns unique MAC addresses to all interfaces assigned to a context.
If you manually assign a MAC address and also enable auto-generation, then the manually assigned MAC address is used. If you later remove the manual MAC address, the auto-generated address is used, if enabled.
In the rare circumstance that the generated MAC address conflicts with another private MAC address in your network, you can manually set the MAC address for the interface.
Because auto-generated addresses (when using a prefix) start with A2, you cannot start manual MAC addresses with A2 if you also want to use auto-generation.
The ASA generates the MAC address using the following format:
A2xx.yyzz.zzzz
Where xx.yy is a user-defined prefix or an autogenerated prefix based on the last two bytes of the interface MAC address, and zz.zzzz is an internal counter generated by the ASA. For the standby MAC address, the address is identical except that the internal counter is increased by 1.
For an example of how the prefix is used, if you set a prefix of 77, then the ASA converts 77 into the hexadecimal value 004D (yyxx). When used in the MAC address, the prefix is reversed (xxyy) to match the ASA native form:
A24D.00zz.zzzz
For a prefix of 1009 (03F1), the MAC address is:
A2F1.03zz.zzzz
Note |
The MAC address format without a prefix is a legacy version. See the mac-address auto command in the command reference for more information about the legacy format. |
About the MTU
The MTU specifies the maximum frame payload size that the ASA can transmit on a given Ethernet interface. The MTU value is the frame size without Ethernet headers, VLAN tagging, or other overhead. For example, when you set the MTU to 1500, the expected frame size is 1518 bytes including the headers, or 1522 when using VLAN. Do not set the MTU value higher to accommodate these headers.
For VXLAN or Geneve, the entire Ethernet datagram is being encapsulated, so the new IP packet is larger and requires a larger MTU: you should set the ASA VTEP source interface MTU to be the network MTU + 54 bytes (for VXLAN) or + 306 bytes (Geneve).
Path MTU Discovery
The ASA supports Path MTU Discovery (as defined in RFC 1191), which lets all devices in a network path between two hosts coordinate the MTU so they can standardize on the lowest MTU in the path.
Default MTU
The default MTU on the ASA is 1500 bytes. This value does not include the 18-22 bytes for the Ethernet header, VLAN tagging, or other overhead.
When you enable VXLAN on the VTEP source interface, if the MTU is less than 1554 bytes, then the ASA automatically raises the MTU to 1554 bytes. In this case, the entire Ethernet datagram is being encapsulated, so the new packet is larger and requires a larger MTU. In general, you should set the ASA source interface MTU to be the network MTU + 54 bytes.
MTU and Fragmentation
For IPv4, if an outgoing IP packet is larger than the specified MTU, it is fragmented into 2 or more frames. Fragments are reassembled at the destination (and sometimes at intermediate hops), and fragmentation can cause performance degradation. For IPv6, packets are typically not allowed to be fragmented at all. Therefore, your IP packets should fit within the MTU size to avoid fragmentation.
For TCP packets, the endpoints typically use their MTU to determine the TCP maximum segment size (MTU - 40, for example). If additional TCP headers are added along the way, for example for site-to-site VPN tunnels, then the TCP MSS might need to be adjusted down by the tunneling entity. See About the TCP MSS.
For UDP or ICMP, the application should take the MTU into account to avoid fragmentation.
Note |
The ASA can receive frames larger than the configured MTU as long as there is room in memory. |
MTU and Jumbo Frames
A larger MTU lets you send larger packets. Larger packets might be more efficient for your network. See the following guidelines:
-
Matching MTUs on the traffic path—We recommend that you set the MTU on all ASA interfaces and other device interfaces along the traffic path to be the same. Matching MTUs prevents intermediate devices from fragmenting the packets.
-
Accommodating jumbo frames—You can set the MTU 9000 bytes or higher when you enable jumbo frames. The maximum depends on the model.
About the TCP MSS
The TCP maximum segment size (MSS) is the size of the TCP payload before any TCP and IP headers are added. UDP packets are not affected. The client and the server exchange TCP MSS values during the three-way handshake when establishing the connection.
You can set the TCP MSS on the ASA for through traffic; by default, the maximum TCP MSS is set to 1380 bytes. This setting is useful when the ASA needs to add to the size of the packet for IPsec VPN encapsulation. However, for non-IPsec endpoints, you should disable the maximum TCP MSS on the ASA.
If you set a maximum TCP MSS, if either endpoint of a connection requests a TCP MSS that is larger than the value set on the ASA, then the ASA overwrites the TCP MSS in the request packet with the ASA maximum. If the host or server does not request a TCP MSS, then the ASA assumes the RFC 793-default value of 536 bytes (IPv4) or 1220 bytes (IPv6), but does not modify the packet. For example, you leave the default MTU as 1500 bytes. A host requests an MSS of 1500 minus the TCP and IP header length, which sets the MSS to 1460. If the ASA maximum TCP MSS is 1380 (the default), then the ASA changes the MSS value in the TCP request packet to 1380. The server then sends packets with 1380-byte payloads. The ASA can then add up to 120 bytes of headers to the packet and still fit in the MTU size of 1500.
You can also configure the minimum TCP MSS; if a host or server requests a very small TCP MSS, the ASA can adjust the value up. By default, the minimum TCP MSS is not enabled.
For to-the-box traffic, including for SSL VPN connections, this setting does not apply. The ASA uses the MTU to derive the TCP MSS: MTU - 40 (IPv4) or MTU - 60 (IPv6).
Default TCP MSS
By default, the maximum TCP MSS on the ASA is 1380 bytes. This default accommodates IPv4 IPsec VPN connections where the headers can equal up to 120 bytes; this value fits within the default MTU of 1500 bytes.
Suggested Maximum TCP MSS Setting
The default TCP MSS assumes the ASA acts as an IPv4 IPsec VPN endpoint and has an MTU of 1500. When the ASA acts as an IPv4 IPsec VPN endpoint, it needs to accommodate up to 120 bytes for TCP and IP headers.
If you change the MTU value, use IPv6, or do not use the ASA as an IPsec VPN endpoint, then you should change the TCP MSS setting.
See the following guidelines:
-
Normal traffic—Disable the TCP MSS limit and accept the value established between connection endpoints. Because connection endpoints typically derive the TCP MSS from the MTU, non-IPsec packets usually fit this TCP MSS.
-
IPv4 IPsec endpoint traffic—Set the maximum TCP MSS to the MTU - 120. For example, if you use jumbo frames and set the MTU to 9000, then you need to set the TCP MSS to 8880 to take advantage of the new MTU.
-
IPv6 IPsec endpoint traffic—Set the maximum TCP MSS to the MTU - 140.
Inter-Interface Communication
Allowing interfaces on the same security level to communicate with each other provides the following benefits:
-
You can configure more than 101 communicating interfaces.
If you use different levels for each interface and do not assign any interfaces to the same security level, you can configure only one interface per level (0 to 100).
-
You want traffic to flow freely between all same security interfaces without ACLs.
If you enable same security interface communication, you can still configure interfaces at different security levels as usual.
Intra-Interface Communication (Routed Firewall Mode)
Intra-interface communication might be useful for VPN traffic that enters an interface, but is then routed out the same interface. The VPN traffic might be unencrypted in this case, or it might be reencrypted for another VPN connection. For example, if you have a hub and spoke VPN network, where the ASA is the hub, and remote VPN networks are spokes, for one spoke to communicate with another spoke, traffic must go into the ASA and then out again to the other spoke.
Note |
All traffic allowed by this feature is still subject to firewall rules. Be careful not to create an asymmetric routing situation that can cause return traffic not to traverse the ASA. |