Ethernet Frame Efficiency Calculator

Introduction:

An Ethernet link carries more than application data. Every frame needs addressing and type fields, a frame check sequence, and sometimes one or more virtual LAN (VLAN) tags. The physical link also spends time on the preamble, start-frame delimiter, and the quiet interval before another frame can begin.

Payload efficiency is the useful payload divided by all bytes charged to one frame time. It explains why a 1 Gbps link cannot deliver 1 Gbps of payload when traffic is made of finite frames. Large payloads spread fixed overhead across more data; short packets repeat the same fixed fields much more often.

The distinction matters in capacity planning. Omitting the 20 wire-time bytes makes efficiency look better than the physical link can achieve. Adding a VLAN tag costs four bytes per frame, while payloads below the minimum data field require padding. Both effects are proportionally larger for small packets.

This is a frame-level ceiling, not an application-throughput promise. IP, transport, tunneling, encryption, retransmissions, congestion, device processing, and protocol-specific physical-layer encoding can all reduce delivered application rate further. Keep the accounting boundary consistent when comparing designs.

How to Use This Tool:

Choose the frame-accounting boundary first, because every efficiency and goodput result depends on it.

1. Enter the useful Payload bytes and the path's Link rate with the correct Kbps, Mbps, Gbps, or Tbps unit.
2. Select a Frame profile for untagged, 802.1Q, QinQ, or MAC-only accounting. Use Custom only when you need 0 to 8 tags or want to exclude wire-time spacing deliberately.
3. Optionally enter a Target payload rate. A zero target leaves capacity checking inactive; a positive target produces required wire rate and headroom.
4. Read Frame ledger for the byte breakdown, then compare the target status with the payload goodput rather than the nominal link rate.

Interpreting Results:

Payload efficiency is a ratio for the selected frame size and accounting boundary. Payload goodput applies that ratio to the nominal link rate, while Frames per second shows how often the frame must be transmitted at full utilization.

• A target fits when payload goodput minus target payload rate is zero or positive. Zero headroom is an exact boundary, not an operating reserve.
• Compare profiles only with the same payload and link rate. A MAC-only result should not be compared directly with a wire-time result.
• A high frame-level efficiency does not prove end-to-end application capacity. Verify packet sizes and actual throughput on the intended path.

Technical Details:

Ethernet frame accounting begins with a 14-byte untagged MAC header. Each VLAN tag adds 4 bytes, and the frame check sequence (FCS) adds 4 bytes. Padding raises the MAC frame to at least 64 bytes. Wire-time mode then adds 8 byte-times for the preamble and start-frame delimiter plus 12 byte-times for the inter-frame gap.

Formula Core:

The equations first resolve the chargeable byte count, then turn it into efficiency, goodput, frame rate, and target capacity.

t is the VLAN-tag count, H is header bytes, P is payload bytes, M is MAC-frame bytes, and g is 20 bytes in wire-time mode or zero in MAC-only mode. W is chargeable bytes, η is efficiency, L is link rate in bits per second, G is payload goodput, and f is frames per second.

A target payload rate T requires T ÷ η bits per second of modeled wire capacity. Headroom is G − T, and the target fits when that value is at least zero. Rates use decimal multipliers: Kbps is 10³ bps, Mbps is 10⁶, Gbps is 10⁹, and Tbps is 10¹². Display formatting rounds values for readability; the comparisons use unrounded numbers.

Limitations:

The model assumes continuous transmission at the entered MAC link rate and applies one frame shape repeatedly. It is most useful for comparing packet-size and tagging choices, not predicting a complete protocol stack.

• Higher-layer headers, encapsulation, packet loss, retransmission, congestion, and device limits are outside the result.
• Physical coding overhead and speed-specific signaling behavior beyond the modeled preamble and gap are not added.
• Payloads up to 9,600 bytes can be modeled, but actual jumbo-frame support and maximum transmission unit limits must be verified end to end.
• All inputs and calculations stay in the browser; exported results contain the values you chose to save.

Worked Examples:

References:

• IEEE 802.3-2022 Standard for Ethernet, IEEE Standards Association, 2022.
• IEEE 802.1Q-2022 Bridges and Bridged Networks, IEEE Standards Association, 2022.
• How to change MTU with ip link, Simplified Guide.