What is the packet overhead for a VXLAN-encapsulated frame?

VXLAN encapsulation adds a series of headers to the original Layer 2 frame to transport it over a Layer 3 network. This added overhead is calculated by summing the sizes of the new headers required for transport across the underlay network. The standard overhead for an IPv4 underlay consists of an Outer Ethernet header (14 bytes), an Outer IP header (20 bytes), an Outer UDP header (8 bytes), and the VXLAN header itself (8 bytes). The total overhead is 14 + 20 + 8 + 8 = 50 bytes. This is a critical consideration for network design, especially regarding MTU settings.

A. 10 bytes: This value is too low and does not account for the multiple headers (L2, L3, L4, and VXLAN) required for encapsulation.

B. 30 bytes: This is an incorrect calculation and does not represent the sum of the standard VXLAN encapsulation headers.

D. 70 bytes: This value represents the overhead when using an IPv6 underlay (40-byte header) instead of IPv4 (14 + 40 + 8 + 8 = 70 bytes). While possible, 50 bytes is the standard for IPv4.

1. Arista Networks, EOS-4.28.0F Manual, "VXLAN Bridging" section, "VXLAN MTU" subsection.

"VXLAN adds 50 bytes of overhead to a packet (14 byte MAC header, 20 byte IPv4 header, 8 byte UDP header, and 8 byte VXLAN header)." This official documentation directly confirms the 50-byte calculation.

2. IETF RFC 7348, "Virtual eXtensible Local Area Network (VXLAN)," Section 5, "VXLAN Packet Format."

This document defines the 8-byte VXLAN header and specifies the encapsulation method: "VXLAN is a Layer 2 overlay scheme over a Layer 3 network. It uses MAC-in-UDP encapsulation..." This supports the inclusion of UDP and underlying IP headers in the overhead calculation.