UDP Throughput Calculator

Calculator inputs

Input mode

Choose whether you know PPS or line rate.

Packets per second (PPS)

Used when your traffic generator is PPS-based.

Link rate (Mbps)

Calculator derives PPS from on-wire bytes.

UDP payload per packet (bytes)

This is your application data (L7).

IP version

Affects the Layer-3 header size.

IPv4/extension options (bytes)

Optional extra bytes beyond base header.

Packet loss (%)

Applied to delivered payload goodput only.

Measurement duration (seconds)

Used to estimate totals transferred.

Rounding (decimal places)

Controls display precision in results.

Layer-2 framing options

Include Ethernet header and FCS

Enables typical Ethernet overhead modeling.

VLAN tag bytes

Adds to the L2 framing bytes if enabled.

Model preamble + IFG

Adds 8 B preamble+SFD and 12 B IFG.

Ethernet header bytes

Typical: 14 bytes (no VLAN).

Ethernet FCS bytes

Typical: 4 bytes FCS.

Extra L2 overhead (bytes)

Use for tunnels, tags, or custom framing.

Layer-4 options

UDP header bytes

Typical UDP header: 8 bytes.

Tip: When modeling Ethernet, keep payload small for PPS-heavy tests.

Example input scenarios

Scenario	Mode	Payload (B)	IPv	Ethernet	Loss (%)	Typical use
High PPS telemetry	PPS = 50,000	200	IPv4	Yes + IFG	0	Sensor bursts and small datagrams
Video over UDP	Link = 100 Mbps	1200	IPv4	Yes + IFG	0.5	Streaming with light loss budget
IPv6 testbed	Link = 1000 Mbps	1400	IPv6	Yes + VLAN	0	Capacity planning for modern networks

These examples are illustrative; use your real packet sizes and rates.

Formula used

On-wire bytes per packet

wire_bytes = payload + UDP + IP + L2_framing

L2_framing can include Ethernet header, FCS, VLAN, extra overhead, and optionally preamble/IFG.

Rates

wire_bps = wire_bytes × 8 × PPS
goodput_bps = payload × 8 × PPS × (1 − loss)

Efficiency is computed as: goodput_bps / wire_bps.

Note: When using “Link rate (Mbps)” mode, the calculator derives PPS by dividing the link rate by the modeled on-wire packet size.

How to use this calculator

Pick an input mode: PPS if you know packet rate, or Link rate if you know bandwidth.
Enter your UDP payload bytes. This is application data per packet.
Choose IPv4/IPv6, and add any extra header bytes if your setup uses them.
Enable Ethernet framing to estimate true on-wire usage, then add VLAN/IFG if needed.
Press Submit. Review goodput, on-wire rate, efficiency, and totals. Use CSV/PDF buttons to export.

Payload, datagram, and on-wire rates

This calculator separates three rates: payload offered, delivered goodput, and on‑wire bandwidth. For example, 1,200‑byte payload at 50,000 PPS offers 480 Mbps of application data. After packet loss, delivered goodput decreases, while on‑wire bandwidth also includes headers and framing.

Header overhead grows at small packet sizes

With IPv4 (20 B) and UDP (8 B), a 200‑byte payload carries 28 bytes of L3/L4 overhead before any L2 framing. That is 12.3% overhead at the IP/UDP layer alone. If you switch to 1,400‑byte payload, the same 28 bytes become just 2.0%.

Ethernet framing and IFG affect true line usage

When Ethernet is enabled, the model can add 14‑byte header, 4‑byte FCS, optional 4‑byte VLAN tag, plus 8‑byte preamble/SFD and 12‑byte inter‑frame gap. Those 20 bytes of preamble+IFG matter most at high PPS. At 100,000 PPS, they add 16 Mbps by themselves.

Link‑rate mode derives PPS from packet size

If you know bandwidth but not PPS, link‑rate mode estimates PPS = link_bps / (wire_bytes×8). A 100 Mbps link with 1,200‑byte payload and Ethernet+IFG typically yields lower PPS than the same link with 200‑byte payload, because each packet occupies fewer bits on the wire.

Loss converts “offered” traffic into delivered goodput

Packet loss is applied to payload delivery. A 0.5% loss on 480 Mbps offered reduces goodput by about 2.4 Mbps, assuming constant PPS and independent loss. This is useful for planning real‑time UDP streams where minor loss is acceptable but jitter buffers are limited.

Efficiency helps capacity planning and sizing

Efficiency = goodput / on‑wire rate summarizes how much bandwidth reaches the application. If efficiency is 85%, a target of 500 Mbps delivered implies roughly 588 Mbps of wire capacity. Use this alongside duration totals to estimate data transfer volume over test windows and maintenance periods.

FAQs

1) What is “goodput” in this tool?

Goodput is delivered UDP payload rate after packet loss. It excludes headers and framing, so it reflects application data that arrives successfully at the receiver.

2) Why is on‑wire rate higher than payload offered?

On‑wire rate includes UDP/IP headers and optional L2 overhead such as Ethernet header, FCS, VLAN, and preamble/IFG. Smaller payloads increase the header share.

3) How should I choose payload size?

Use the application’s real datagram size. For path MTU safety, many systems choose about 1,200 bytes payload on IPv4/IPv6 to reduce fragmentation risk.

4) Does this model account for retransmissions?

No. UDP has no built‑in retransmission. If your application retries at a higher layer, treat that as higher PPS or higher offered rate and re‑calculate.

5) What does “IP options bytes” represent?

It represents extra IP header bytes beyond the base header. In most production networks it is zero, but some lab setups model additional header extensions.

6) Why does link‑rate mode change the calculated PPS?

Because the same bandwidth can carry many small packets or fewer large packets. The calculator divides link rate by modeled bits per packet to estimate PPS.