Computer Science · Network Engineering · Protocol Analysis
Bandwidth and Throughput Calculator
Calculates effective network throughput, transfer time, and efficiency given bandwidth, overhead, and payload size.
Calculator
Formula
Throughput is the effective data rate achieved over a channel. Bandwidth is the raw link capacity in bits per second. \(\eta\) (eta) is the protocol efficiency, where \(L_{\text{payload}}\) is the useful payload size per frame or packet and \(L_{\text{overhead}}\) is the header/trailer overhead per frame. \(T_{\text{transfer}}\) is the estimated time to transfer a file of size \(S_{\text{file}}\), computed as file size divided by throughput.
Source: Forouzan, B.A. — Data Communications and Networking, 5th Edition, McGraw-Hill; IEEE 802 Standards Series.
How it works
Bandwidth and throughput are often used interchangeably, but they represent distinct concepts. Bandwidth is the theoretical maximum capacity of a network link — the raw bit rate the physical medium can carry, expressed in bits per second (bps) or megabits per second (Mbps). Throughput is the actual rate of useful data delivered to an application. The difference arises because every networking protocol adds overhead: headers, trailers, checksums, and control fields that consume link capacity without carrying application payload.
The key formula is: Throughput = Bandwidth × η, where η (eta) is the protocol efficiency defined as η = Lpayload / (Lpayload + Loverhead). For example, a TCP/IP segment over Ethernet carries a 1460-byte payload but includes 20 bytes of IP header, 20 bytes of TCP header, and 14 bytes of Ethernet frame header plus 4-byte FCS — totalling 58 bytes of overhead. This pushes efficiency to roughly 1460 / 1518 ≈ 0.962, meaning about 3.8% of link capacity is consumed by protocol overhead alone. Real-world efficiency is often lower still, due to retransmissions, ACK traffic, inter-frame gaps, and application-layer framing.
Once throughput is known, transfer time is simply T = Sfile / Throughput, where file size is converted to bits. This estimate is ideal for planning bulk data migrations, backup windows, streaming pipelines, and cloud storage uploads. Network professionals also use this calculation to validate whether a proposed WAN link can meet Service Level Agreements (SLAs), and data engineers use it to size ingestion pipelines for large dataset transfers between clusters or data lakes.
Worked example
Suppose you are planning a file transfer over a 100 Mbps Ethernet link and want to know how long it will take to move a 500 MB database backup. The application uses standard TCP/IP, with each frame carrying a payload of 1460 bytes and an overhead of 60 bytes (20 bytes IP header + 20 bytes TCP header + 14 bytes Ethernet header + 4 bytes FCS + 2 bytes inter-frame rounding).
Step 1 — Calculate Protocol Efficiency:
η = 1460 / (1460 + 60) = 1460 / 1520 ≈ 0.9605
Step 2 — Calculate Effective Throughput:
Throughput = 100 Mbps × 0.9605 = 96.05 Mbps
In MB/s: 96.05 / 8 = 12.006 MB/s
Step 3 — Calculate Transfer Time:
T = (500 MB × 8 bits/byte) / 96.05 Mbps = 4000 Mb / 96.05 Mbps ≈ 41.64 seconds
This means the 500 MB backup will complete in just under 42 seconds under ideal conditions. In practice, add margin for TCP slow start, network congestion, disk I/O bottlenecks, and encryption overhead if TLS is in use.
Limitations & notes
This calculator models a single, ideal link with no queuing delay, retransmission, or congestion. Real throughput is reduced by factors including TCP congestion control and slow start, round-trip latency (especially significant for high-bandwidth, high-latency WAN links), packet loss and retransmission, full-duplex vs. half-duplex operation, shared medium contention on wireless or hub-based networks, and application-layer encryption overhead (e.g., TLS adds ~5–10% CPU and framing cost). The bandwidth-delay product is a critical concept for long-distance links — a 100 Mbps link with 200 ms RTT requires a TCP window size of at least 2.5 MB to stay fully utilized, and default OS buffers may limit actual throughput well below the theoretical maximum. For WAN or satellite links, use the Bandwidth-Delay Product formula alongside this tool for accurate capacity planning.
Frequently asked questions
What is the difference between bandwidth and throughput?
Bandwidth is the theoretical maximum capacity of a network link — the raw bit rate of the physical medium. Throughput is the actual rate of useful data delivered end-to-end, after subtracting protocol overhead, retransmissions, and other inefficiencies. Throughput is always less than or equal to bandwidth.
Why is my throughput much lower than my advertised bandwidth?
Several factors reduce throughput below the nominal bandwidth: protocol headers consume capacity, TCP's congestion control and slow-start mechanism limit early transfer rates, network latency constrains the maximum window size, and shared infrastructure introduces contention. For long-distance links, the bandwidth-delay product determines the maximum achievable throughput for a single TCP stream.
How do I convert Mbps to MB/s?
Divide Mbps by 8 to get megabytes per second (MB/s), because there are 8 bits in a byte. For example, 100 Mbps ÷ 8 = 12.5 MB/s. Note that ISPs typically advertise speeds in megabits per second, while file managers display transfer rates in megabytes per second, which causes common confusion.
What overhead values should I use for TCP/IP over Ethernet?
A standard TCP/IP over Ethernet frame has 14 bytes of Ethernet header, 4 bytes of FCS, 20 bytes of IPv4 header, and 20 bytes of TCP header — totalling 58 bytes of overhead per 1460-byte payload segment. If IPv6 is used, the IP header is 40 bytes, increasing overhead to 78 bytes. TLS adds a further 13-byte record header plus padding.
Can this calculator be used for wireless (Wi-Fi) networks?
The formula is applicable, but Wi-Fi introduces additional efficiency losses beyond header overhead: CSMA/CA collision avoidance, OFDM guard intervals, interframe spaces (DIFS, SIFS), and the MAC layer acknowledgement mechanism. Typical Wi-Fi efficiency ranges from 50–70% of the PHY rate, so real throughput on an 802.11ac 433 Mbps link may only reach 200–300 Mbps under good conditions.
Last updated: 2025-01-15 · Formula verified against primary sources.