Fortinet black logo

Administration Guide

VXLAN

VXLAN

Virtual Extensible LAN (VXLAN) is a network virtualization technology used in large cloud computing deployments. It encapsulates layer 2 Ethernet frames within layer 3 IP packets using the UDP transport protocol on port 4789. VXLAN endpoints that terminate VXLAN tunnels can be virtual or physical switch ports, and are known as VXLAN tunnel endpoints (VTEPs).

Sample VXLAN packet

A VXLAN packet encapsulation occurs by first inserting a VXLAN header in front of the original layer 2 frame. This VXLAN header uses 3 B for the VNID that is used to identify the VXLAN segment, meaning that there are 16,777,215 different possible VNIDs. This allows for more unique LAN segments than possible VLANs. The original frame and the VXLAN header are then encapsulated into the UDP payload. The outer IP header allows it to be routed and transported over a layer 3 network, thus providing a layer 2 overlay scheme over a layer 3 network.

This equates to 50 B of overhead over the original frame: 14 B (Ethernet) + 20 B (IPv4) + 8 B (UDP) + 8 B (VXLAN headers). Since fragmenting a VXLAN packet is not recommended, it is advisable to increase the MTU size to 1550 B or above if possible, or to decrease the TCP MSS size inside a firewall policy.

For more information about VXLAN, see RFC 7348.

The following topics provide information about VXLAN:

VXLAN

Virtual Extensible LAN (VXLAN) is a network virtualization technology used in large cloud computing deployments. It encapsulates layer 2 Ethernet frames within layer 3 IP packets using the UDP transport protocol on port 4789. VXLAN endpoints that terminate VXLAN tunnels can be virtual or physical switch ports, and are known as VXLAN tunnel endpoints (VTEPs).

Sample VXLAN packet

A VXLAN packet encapsulation occurs by first inserting a VXLAN header in front of the original layer 2 frame. This VXLAN header uses 3 B for the VNID that is used to identify the VXLAN segment, meaning that there are 16,777,215 different possible VNIDs. This allows for more unique LAN segments than possible VLANs. The original frame and the VXLAN header are then encapsulated into the UDP payload. The outer IP header allows it to be routed and transported over a layer 3 network, thus providing a layer 2 overlay scheme over a layer 3 network.

This equates to 50 B of overhead over the original frame: 14 B (Ethernet) + 20 B (IPv4) + 8 B (UDP) + 8 B (VXLAN headers). Since fragmenting a VXLAN packet is not recommended, it is advisable to increase the MTU size to 1550 B or above if possible, or to decrease the TCP MSS size inside a firewall policy.

For more information about VXLAN, see RFC 7348.

The following topics provide information about VXLAN: