Managing Egress Flooding

Egress flooding takes action on a packet based on the packet destination MAC address. By default, egress flooding is enabled, and any packet for which the destination address is not in the FDB (forwarding database) is flooded to all ports except the ingress port.

You can enhance security and privacy as well as improve network performance by disabling Layer 2 egress flooding on a port, VLAN (Virtual LAN), or VMAN. This is particularly useful when you are working on an edge device in the network. Limiting flooded egress packets to selected interfaces is also known as upstream forwarding.

Note

Note

Disabling egress flooding can affect many protocols, such as IP and ARP.

Upstream Forwarding or Disabling Egress Flooding Example illustrates a case where you want to disable Layer 2 egress flooding on specified ports to enhance security and network performance.

Click to expand in new window
Upstream Forwarding or Disabling Egress Flooding Example
GUID-E17B6307-A2C0-4F25-A55B-60DD53292616-low.png

In this example, the three ports are in an ISP-access VLAN. Ports 1 and 2 are connected to clients 1 and 2, respectively, and port 3 is an uplink to the ISP network. Because clients 1 and 2 are in the same VLAN, client 1 could possibly learn about the other client‘s traffic by sniffing client 2‘s broadcast traffic; client 1 could then possibly launch an attack on client 2.

However, when you disable all egress flooding on ports 1 and 2, this sort of attack is impossible, for the following reasons:
  • Broadcast and multicast traffic from the clients is forwarded only to the uplink port.
  • Any packet with unlearned destination MAC addresses is forwarded only to the uplink port.
  • One client cannot learn any information from the other client. Because egress flooding is disabled on the access ports, the only packets forwarded to each access port are those packets that are specifically targeted for one of the ports. There is no traffic leakage.

In this way, the communication between client 1 and client 2 is controlled. If client 1 needs to communicate with client 2 and has that IP address, client 1 sends out an ARP request to resolve the IP address for client 2.