Ethernet Serialization Delay Calculator
Serialization delay is the time a port needs to clock a frame onto the wire: frame bits ÷ line rate. A 1,518-byte frame takes 12.14 µs at 1 Gbps but 121.4 ns at 100 Gbps. Each frame also occupies 20 extra bytes of preamble, SFD and inter-frame gap, which sets the famous 14.88 Mpps line rate for 64-byte frames at 10 Gbps. This calculator gives per-hop and multi-hop delay, wire time, frames per second and L2 efficiency.
Frame and link
Enter the Ethernet frame size including the 4-byte FCS, the line rate and the number of store-and-forward hops.
64 minimum, 1518 untagged, 1522 with 802.1Q, 9018+ jumbo.
Each store-and-forward hop adds one frame time.
Core calculation
Serialization delay = Frame bits ÷ Link rate. Wire time = (Frame + Preamble/SFD + IFG) bits ÷ rate. Frame rate = rate ÷ bits on the wire. Total delay ≈ hops × per-hop serialization.Serialization. Time to clock one frame onto the wire at the configured link rate.
Wire time. Include preamble/SFD and IFG when estimating media occupancy.
Multi-hop. Sum per-hop serialization for store-and-forward paths; cut-through is lower.
Scope. Propagation, queuing and processing delay are outside this calculator.
Calculation result
Delay = frame bytes × 8 ÷ rate. Frame rate = rate ÷ ((frame + 20) × 8).
Serialization delay per hop
—Total over all hops
—Cut-through switches start forwarding before the whole frame arrives, so they add less than one frame time. Propagation, queuing and processing delay are not included.
Methodology and engineering notes
METHODOLOGY
How to use
- Enter the frame size from destination MAC to FCS (64–1518 bytes for standard frames).
- Pick a line rate or choose Custom.
- Enter how many store-and-forward devices the frame crosses.
- Read per-hop and total delay, wire time, frames per second and L2 efficiency.
Serialization formula
Serialization delay = frame bytes × 8 ÷ line rate. On the wire every frame is preceded by a 7-byte preamble and 1-byte SFD and followed by a minimum 96-bit (12-byte) inter-packet gap, so it occupies frame + 20 bytes. Wire time = (frame + 20) × 8 ÷ rate.
Frame rate and efficiency
Maximum frames per second = rate ÷ ((frame + 20) × 8). L2 efficiency = frame ÷ (frame + 20); L2 throughput = efficiency × rate. 64-byte frames at 10 Gbps give 14,880,952 fps and 76.19% efficiency; 1518-byte frames give 812,743 fps and 98.70%.
Worked case: 64 B and 1518 B at 10 Gbps
A 64-byte frame takes 51.20 ns per hop, 153.60 ns over three store-and-forward hops, and occupies 84 bytes (67.20 ns) on the wire: 14,880,952 fps. A 1518-byte frame takes 1.214 µs per hop and 3.643 µs over three hops; at 1 Gbps the same frame takes 12.144 µs per hop.
Frequently asked questions
FAQ
How do I calculate serialization delay?
Divide the frame size in bits by the line rate. 1500 bytes × 8 = 12,000 bits; at 10 Gbps that is 1.2 µs, at 1 Gbps 12 µs.
Why is 10G line rate 14.88 Mpps?
A minimum 64-byte frame occupies 84 bytes with preamble, SFD and IFG. 10,000,000,000 ÷ (84 × 8) = 14,880,952 frames per second.
Does serialization delay depend on cable length?
No. It depends only on frame size and line rate. Cable length adds propagation delay, about 5 µs per km of fiber.
What is the overhead of an Ethernet frame on the wire?
20 bytes per frame: 7 bytes preamble, 1 byte SFD and 12 bytes minimum inter-packet gap, in addition to the 18-byte MAC header and FCS inside the frame.
Do jumbo frames reduce latency?
They raise efficiency but increase per-frame serialization delay: a 9018-byte frame takes 7.2 µs at 10 Gbps versus 1.2 µs for 1518 bytes.
Recent calculations
PROJECT RECORDS
Standards and references
ENGINEERING REFERENCES
- IEEE 802.3-2022: Clause 3 MAC frame (7-byte preamble, 1-byte SFD, FCS) and Clause 4.4.2 interPacketGap (96 bit times) [clause number UNVERIFIED]
- RFC 2544 (1999): benchmarking methodology – section 9 standard frame sizes 64–1518 bytes
- RFC 1242 (1991): benchmarking terminology – section 3.8 latency of store-and-forward devices
The 20-byte per-frame overhead uses the minimum IFG; some PHYs and rate adaptation (for example WAN PHY) change the effective payload rate. Last reviewed: 2026-10-20.
ENGINEERING WORKFLOW
Continue the engineering workflow
Pick the Spanning Tree forwarding path, then estimate the per-hop serialization delay along it.