Data Transfer Rate Conversion Guide

← Voltar ao conversor de Taxa de Transferência de Dados

Tradução em andamento.

Este guia ainda não foi traduzido para PT. A versão em inglês é mostrada abaixo.

Data transfer is where bits and bytes collide and cost users 8× their expected speeds. A '100 Mbps' fibre plan delivers 12.5 MB/s of file-transfer throughput, not 100 — because Mbps is megabits per second and MB/s is megabytes per second, and one byte is eight bits. This guide covers every transfer-rate unit we support, the exact factors between bits and bytes, and the everyday gotchas: ISP advertising rate vs real throughput, gigabit Ethernet vs gigabyte per second, and the protocol overhead that eats roughly 10% before any data hits your application.

What is data transfer rate?

Data transfer rate is the speed at which information moves through a channel, measured in bits or bytes per unit time. The SI base for our converter is the bit per second (bps), the lowest-common-denominator unit used by network standards: 10BASE-T Ethernet is 10,000,000 bps, gigabit Ethernet is 1,000,000,000 bps, and so on. File-transfer rates are usually quoted in bytes per second (B/s) because file sizes are in bytes; the conversion is a fixed factor of 8.

The split between bits-per-second (used by network and ISP marketing) and bytes-per-second (used by file managers and operating systems) is the single most common source of confusion in transfer-rate questions. Marketing teams choose the bigger-looking number; users choose whichever matches their software. Our converter accepts both on equal footing.

History of data transfer units

The bits-per-second convention traces to early telephone-line modems, where bandwidth was naturally measured at the signalling layer. A 300-baud modem in the 1970s transferred 300 bits per second. By the 1990s, 14.4 and 28.8 'kbps' modems were household standard. ISPs kept the bits-per-second convention because they sell line rate at the network layer, where bits are the unit.

Operating systems and file managers report bytes per second because file sizes are in bytes and the user expectation is 'how long until this download finishes'. The two systems coexist with a fixed 8× factor, and the only thing left to negotiate is whether 'kilo' means 1,000 or 1,024. In data-transfer contexts, the decimal convention is essentially universal — there is no 'kibibits per second' in any standards document — so 1 Mbps = 1,000,000 bits per second exactly, and 1 MB/s = 1,000,000 bytes per second exactly. The decimal-vs-binary battle that complicates digital storage does not extend to transfer rates.

Modern Ethernet and Wi-Fi standards use multiplier suffixes that map cleanly: 100BASE-T = 100 Mbps, gigabit Ethernet = 1 Gbps = 1,000 Mbps, 10 gigabit Ethernet = 10 Gbps. Wi-Fi advertised rates are link rates (the wireless equivalent of line rate), and actual application throughput is typically 30–60% lower because of half-duplex, retransmits, and protocol overhead.

Units we cover

Our data-transfer category supports eight units across two families. Bits per second: bit per second (bps, SI base), kilobit per second (kbps, 1,000 bps), megabit per second (Mbps, 10⁶ bps), gigabit per second (Gbps, 10⁹ bps), terabit per second (Tbps, 10¹² bps). Bytes per second: kilobyte per second (KB/s, 8,000 bps), megabyte per second (MB/s, 8 × 10⁶ bps), gigabyte per second (GB/s, 8 × 10⁹ bps). Each unit has its exact bps equivalent published, and Calcflux applies those factors directly.

All units in this category use decimal prefixes (powers of 1,000), matching network-engineering convention. If you encounter '1 MiB/s' in a Linux tool's output, that's a binary mebibyte per second (1,048,576 B/s = 8,388,608 bps); it differs from MB/s by 4.9%. We do not introduce binary IEC prefixes here because they are essentially absent from networking specifications and would create more confusion than they resolve.

How to convert

Every transfer-rate conversion routes through the bit per second. Each unit declares how many bps it represents: 1 kbps = 1,000, 1 Mbps = 1,000,000, 1 Gbps = 10⁹, 1 KB/s = 8,000, 1 MB/s = 8,000,000, 1 GB/s = 8 × 10⁹. The methodology page lists every factor.

Useful approximations: 1 Gbps = 125 MB/s line rate, ≈ 117 MB/s after typical Ethernet and TCP overhead. 100 Mbps = 12.5 MB/s line rate, ≈ 11–12 MB/s real-world. 1 Mbps = 125 kB/s. The protocol-overhead factor sits at roughly 90–95% for tuned TCP over Ethernet, lower for Wi-Fi or congested links.

Streaming and download anchors are worth keeping in mind. Netflix HD streaming uses about 5 Mbps; 4K streaming uses about 25 Mbps; an 8K stream uses 80–100 Mbps. A standard-definition video call uses 0.5 Mbps in each direction. Downloading a 5 GB game over a 100 Mbps connection takes at least 400 seconds (6.7 minutes) of wire time, plus protocol overhead. Cloud-backup software typically caps at 1–2 MB/s per stream to avoid saturating the link, which is the right ballpark for residential upload plans rated at 10–20 Mbps.

Common conversions you'll need

ISP plans, file downloads, and network sizing drive most data-transfer searches. Megabits per second to megabytes per second handles the everyday 'why is my download slower than my plan' question — divide by 8 to translate. Gigabits per second to megabytes per second sizes Gigabit Ethernet against real file transfers: 1 Gbps maxes out at 125 MB/s before any protocol overhead. Kilobits per second to megabits per second covers legacy modem-era rates and IoT links — a 64 kbps voice-over-IP channel is 0.064 Mbps. Gigabits per second to gigabytes per second handles 10/40/100 GbE backbone planning: 100 GbE is 12.5 GB/s line rate.

The full data-transfer comparison table lists every supported unit with its bps factor and a worked sample.

Pitfalls and gotchas

Bits vs bytes: 1 byte = 8 bits, so Mbps and MB/s differ by a factor of 8. A '300 Mbps' Wi-Fi plan caps file transfers at 37.5 MB/s — not 300 MB/s. The factor of 8 is by far the most common cause of disappointed download-speed expectations.

Line rate vs application throughput: 1 Gbps Ethernet is 125 MB/s of raw line rate. After Ethernet preambles, inter-frame gaps, TCP/IP headers, and ACKs, real-world TCP throughput on a tuned connection is typically 117–119 MB/s — about 6% overhead. Add an MTU below 1500 or Wi-Fi half-duplex and the gap widens further. Vendor benchmarks at '950 Mbps' for gigabit are real, not exaggeration.

Mbps vs MB/s case sensitivity: 'Mbps' (lowercase b) is megabits per second; 'MB/s' (capital B) is megabytes per second. Some style guides write 'Mbit/s' to remove the ambiguity, but it has not caught on. Read carefully — Apple and Microsoft both sometimes flip the convention on the same product page.

Wi-Fi rated speeds: 'Wi-Fi 6 AX3000' advertises 3 Gbps of theoretical aggregate link rate across all spatial streams. A single device on a single channel sees a fraction of that, and after half-duplex and protocol overhead, application throughput is typically 30–60% of the link rate. Treat AX numbers as ceiling values, not delivered speeds.

Ethernet line rate vs gigabytes per second: gigabit Ethernet is 1 Gbps = 0.125 GB/s, not 1 GB/s. Confusing the two by an order of magnitude can lead to dramatically undersized backbone plans. 10 GbE delivers about 1.25 GB/s; 100 GbE delivers about 12.5 GB/s; 'gigabyte-per-second backbone' typically implies 10 GbE or better.