Data Transfer Rate Converter

A data transfer rate converter takes any number expressed in one rate unit, bit/s, kbit/s, Mbit/s, Gbit/s, Tbit/s, byte/s, kB/s, MB/s, GB/s, KiB/s, MiB/s, or GiB/s, and expresses it in the other eleven, all at once. Every field stays in sync because the converter pivots through one base unit, the bit per second, just as the torque converter pivots through the newton-metre and the pressure converter pivots through the pascal. The pivot is the only place where the bits-vs-bytes distinction (×8) and the SI-vs-IEC distinction (×1000 vs ×1024) are merged: every other field is derived from it, so the factor chain is exact and reversible. A second panel below the rate grid estimates how long a file of a given size takes to transfer at the current rate, which is the most common real-world use case.

This kind of converter shows up in a few places. Network engineers size backhaul links against benchmarks expressed in Mbit/s when the equipment is sold in MB/s. Software developers reading download pages need to convert the rate advertised in the connect dialog into the bytes-per-second figure an OS reports. Storage-system designers compare a SATA link budget (Gbit/s) against the throughput an SSD can sustain (GB/s or GiB/s). Home users ask why their "100 Mbit/s" fibre line never seems to push past ~11 MB/s in a browser download, the answer is the 8× factor, plus overhead.

How to Use the Data Transfer Rate Converter

1. Type a number into any one of the twelve rate fields. The other eleven update instantly, so you can never end up with inconsistent answers across fields.
2. Pick the SI decimal set (k, M, G, T) or the IEC binary set (Ki, Mi, Gi) deliberately: they are labelled in the field name and they differ by a factor of 1.024 per step. The pivot always uses the SI value internally, so a 1 MiB/s rate shows up as 8.388608 Mbit/s, not 8.
3. Use the connection presets for a quick sanity check on a typical real-world tier, ADSL, VDSL/entry fibre, fibre-to-the-home, 4G LTE, gigabit ethernet. Each preset loads the line-rate value into the rate grid.
4. In the Download time panel, type a file size and pick its unit. The panel uses the rate currently in the rate grid to compute the line-rate transfer time. The figures reflect the wire speed; real throughput will be lower because of protocol overhead (see "Where the Numbers Come From" below).
5. To find a missing value: pick the field whose value you do not know, type 0 there, and type the known value into the field whose unit is the most natural (e.g. type into MB/s for a download size; type into Mbit/s for a connection rate).
6. Negative input is treated as zero in the rate grid to avoid producing negative bandwidth numbers, leave the field empty or use 0 if you want a clean reset.

The Formulas: Conversion Through the bit/s Pivot

Every conversion in this tool passes through the bit per second (bit/s), the SI coherent unit for information rate (NIST SP 811). Once a value is expressed in bit/s, dividing by the bit/s factor of any other unit yields the equivalent in that unit. The factors below are exact (or exact to the digits shown) under IEC 80000-13 (binary prefixes), NIST SP 811 (SI decimal prefixes), and IEEE 802.3 (Ethernet octets, 1 octet = 8 bit).

SI decimal bit-based rates (1 k = 1,000):

• 1 bit/s = 1 bit/s (the base)
• 1 kbit/s = 1,000 bit/s
• 1 Mbit/s = 1,000,000 bit/s = 1e6 bit/s
• 1 Gbit/s = 1,000,000,000 bit/s = 1e9 bit/s
• 1 Tbit/s = 1,000,000,000,000 bit/s = 1e12 bit/s

SI decimal byte-based rates (1 byte = 8 bit):

• 1 byte/s = 8 bit/s
• 1 kB/s = 1,000 byte/s = 8,000 bit/s
• 1 MB/s = 1,000,000 byte/s = 8,000,000 bit/s
• 1 GB/s = 1,000,000,000 byte/s = 8,000,000,000 bit/s

IEC binary rates (1 Ki = 1,024):

• 1 KiB/s = 1,024 byte/s = 8,192 bit/s
• 1 MiB/s = 1,048,576 byte/s = 8,388,608 bit/s = 8.388608 Mbit/s
• 1 GiB/s = 1,073,741,824 byte/s = 8,589,934,592 bit/s

Direction A, Convert into bit/s: multiply the source value by its factor above. Example: 12.5 MB/s × 8,000,000 = 100,000,000 bit/s = 100 Mbit/s.

Direction B, Convert out of bit/s: divide the bit/s value by the factor of the target unit. Example: 100,000,000 bit/s ÷ 8,000,000 = 12.5 MB/s.

Download time uses the same pivot. A file-size value in any of B, KB, MB, GB, KiB, MiB, or GiB is first expressed in bytes (and then multiplied by 8 to get bits). The rate value is read in bit/s, and transfer time in seconds is size_bits / rate_bits_per_s. Convert seconds into the largest natural unit: hours if ≥ 3600, minutes if ≥ 60, otherwise seconds.

Why pivot through bit/s? Because the bit is the SI coherent base of information, and 1 byte = 8 bits is exact under IEEE 802.3. Going through any other unit would force a double conversion in one direction and lose precision on the slow half of the chain.

Why are some pairs off by 1.024, others by 8? The 1.024 factor is the binary step (1 Ki = 1,024 = 2^10), which appears only between SI and IEC scales. The 8 factor is the byte step (1 byte = 8 bit), which appears between any bit-based and any byte-based unit. The two steps multiply where they overlap: 1 MiB/s ÷ 1 MB/s = 1.024^2 = 1.048576, and 1 MB/s ÷ 8 bit/s = 1e6, so 1 MiB/s = 8.388608 × (1 MB/s ÷ 1e6 ÷ 8) bit/s.

Worked Examples

Example 1, A 100 Mbit/s fibre link expressed in MB/s. A typical home-fibre plan advertises 100 Mbit/s downstream. The browser or download manager, however, displays the actual throughput in MB/s.

100 × 1,000,000 bit/s ÷ 8 ÷ 1,000,000 = 12.5 MB/s

So a 100 Mbit/s plan has a theoretical max downstream of 12.5 MB/s. Real-world throughput is below 12.5 MB/s in practice; see "Common Mistakes."

Example 2, A 1 GB/s SSD sequential write expressed in Mbit/s and MiB/s. A modern NVMe SSD sustains roughly 1 GB/s sequential writes. The same throughput in Mbit/s is 1 GB/s × 8 = 8,000 Mbit/s = 8 Gbit/s. In IEC binary terms, 1 GB/s ÷ 1.073741824 ≈ 0.931 GiB/s, so the same SSD writes about 0.93 GiB/s in IEC language.

So 1 GB/s = 8,000 Mbit/s = 8 Gbit/s = 0.931 GiB/s.

Example 3, Downloading a 700 MB file at 10 Mbit/s (the classic CD-image scenario). The bits are 700 MB × 8 = 5,600,000,000 bits = 5.6 Gb. Divide by the 10 Mbit/s rate:

5.6 × 10^9 bit ÷ (10 × 10^6 bit/s) = 560 seconds = 9 minutes 20 seconds.

This is the textbook answer, and the worst-case answer on a sustained-rate assumption. Real TCP throughput is lower by tens of percent on average and is much lower than 560 s on a congested link.

Example 4, A 1 Gbit/s link serving a 4.7 GB DVD image. Bits: 4.7 × 10^9 × 8 = 3.76 × 10^10 bit. Divide by 1 × 10^9 bit/s:

3.76 × 10^10 ÷ 10^9 = 37.6 seconds.

So a 4.7 GB DVD image finishes in roughly 37.6 seconds on a saturated 1 Gbit/s link. On a real link the figure is in the 40-to-50 second range depending on protocol overhead and window scaling.

Example 5, A 1 Tbit/s backbone line carrying 125,000 MB/s. A 1 Tbit/s optical carrier is the backbone pipeline for many transit networks. To put that in byte-rate form: 1 Tbit/s ÷ 8 = 125 GB/s = 125,000 MB/s = 8,000,000 MiB/s = approximately 116.4 GiB/s.

So 1 Tbit/s = 125,000 MB/s = 125 GB/s. A 1 Tbit/s line moves roughly 116 GiB/s in IEC binary terms, the 1.024^n slip between MB and MiB is meaningful when volumes are this large.

Where the Numbers Come From (and Why Measured Throughput Is Lower)

The figures shown in the converter are line rates, the number of bits per second the wire or wireless channel is engineered to carry. What a user actually sees is application-layer throughput, which is always lower. The gap is the sum of protocol overhead:

• Physical-layer encoding overhead (e.g. 64b/66b on 10 Gbit+ ethernet: roughly 3% overhead; 8b/10b on 1 Gbit ethernet: 20% overhead).
• Inter-packet gap, preamble, and frame headers (ethernet headers, IP headers, TCP headers, TLS records).
• TCP acknowledgments and retransmissions (recoveries for lost packets, which are especially costly on wireless links).
• TLS handshake size if the connection is HTTPS.
• Slow-start, congestion avoidance, and bufferbloat at bottlenecks.

Together these can take anywhere from a few percent off on a clean fibre link up to well over half on a busy mobile or Wi-Fi link. Exact percentages depend on the path, the protocol, the packet size, and the link state, do not quote a single number. The rule of thumb is "real throughput is somewhere between 50% and 95% of line rate, with most clean wired cases in the 80 to 95% range and most wireless cases in the 50 to 80% range."

The download-time panel takes the line rate as the denominator and ignores overhead by design, this gives you the lower bound on transfer time (the best you can possibly get). When you measure real throughput against the panel, the gap is your overhead.

Common Mistakes

Confusing Mbit/s with MB/s, an 8× error. They are not the same unit, but the prefix letters look almost identical on a printout. A "100 Mbit/s" connection delivers at most 12.5 MB/s because 100 ÷ 8 = 12.5 (and the rest is overhead). A misread of "100 MB/s" instead of "100 Mbit/s" gives you an 8× over-estimate of connection speed, or an 8× under-estimate of download time.

Confusing MB (SI, 10^6) with MiB (IEC, 2^20), a 4.86% error compounded by step. 1 MB = 1,000,000 bytes; 1 MiB = 1,048,576 bytes. The factor is 1.048576 per step, so 1 GB ÷ 1 GiB = 0.931. Operating systems often reported file sizes in "MB" but actually meant MiB; modern OSes have migrated to GiB/MiB unit labels but the confusion persists in download tools. The rule: 1 TB drive labelled in SI bytes shows roughly 931 GB when the OS uses IEC binary. The reason is that drive makers advertise in SI (1 TB = 1,000,000,000,000 bytes), but most operating systems and some applications historically reported disk usage in IEC (1 TiB = 1,099,511,627,776 bytes), so a 1 TB drive has only 1,000,000,000,000 / 1,073,741,824 = 931.32 GiB of addressable space.

Reading "line rate" as a guarantee of throughput. A 100 Mbit/s plan will never push 12.5 MB/s in real browsing. Protocol overhead, congestion control, and link contention trim the actual figure by anywhere from small fractions up to significant percentages. Use the download-time panel for the lower bound; multiply by 1.1 to 2.0 in your head for a realistic estimate on a clean fibre link, and larger on a wireless or mobile link.

Computing download time without converting the file size. A "700 MB" image is 700 MB of bytes regardless of the connection's rate, but the user often quotes the connection in Mbit/s. The 8× error in size and the 8× error in rate cancel only if both are on the same scale. Be explicit: 700 × 8 = 5,600 megabits of payload, and at 10 Mbit/s = 560 s = 9 min 20 s.

Treating latency and bandwidth as the same metric. Bandwidth is the number of bits per second; latency is the time each bit takes to traverse the link. They are independent. Throughput at the application layer is bounded by min(bandwidth, window_size / round_trip_time) for window-limited protocols like TCP, so a gigabit link with a 200 ms round-trip and a 64 KB window never delivers more than 64 KB / 0.2 s = 320 KB/s until the window scales. Fast links across long distances are bandwidth-rich but latency-bound; this is why CDNs, edge caches, and TCP window scaling exist.

Frequently Asked Questions

What is the difference between bit/s and byte/s?

A bit is one binary digit, a 0 or a 1. A byte is a group of 8 bits (an "octet" in IEEE 802.3 networking). Network lines transmit bits one at a time, so connection rates are quoted in bit/s and its multiples. Files are stored in bytes, so download tools and operating systems report speeds in byte/s. The ratio is exactly 8 bit per byte: 8 Mbit/s = 1 MB/s.

Why do ISPs advertise speeds in megabits per second instead of megabytes per second?

Because the bit is the natural unit of the wire, modem rates, optical carrier rates, and wireless channel capacities are all measured at the bit level. Historical reasons reinforce the convention: the earliest digital modems reported their symbol rate, not their byte rate, and the SI prefixes (k, M, G) come from the bit side. Operating systems, browsers, and download managers now report bytes per second for human convenience. The mismatch is one of the most common sources of customer confusion about service speed.

What is the difference between MB and MiB?

MB is the SI decimal megabyte: 1 MB = 1,000,000 bytes (10^6). MiB is the IEC binary mebibyte: 1 MiB = 1,048,576 bytes (2^20). The ratio is 1.048576. IEC 80000-13 introduced the binary prefixes (Ki, Mi, Gi, Ti, Pi, Ei) precisely to remove the ambiguity that "MB" could mean either. In practice, "MB" in storage contexts usually means MiB on older systems and MB on modern ones.

Why does a 1 TB drive show up as 931 GB?

Drive makers use SI decimal units (1 TB = 10^12 bytes). Most operating systems historically used IEC binary units (1 TiB = 2^40 bytes) when reporting disk usage, so a 1 TB drive has 10^12 ÷ 2^40 = 931.32 GiB of addressable space. Newer OSes label storage correctly with GiB/TiB; the discrepancy is rooted in decades of mixed-unit usage that IEC 80000-13 set out to fix.

Does this calculator account for real-world protocol overhead?

No, it computes the line-rate figure (the lower bound on transfer time). Real TCP/IP throughput is reduced by physical-layer encoding, frame headers, ACKs, congestion control, TLS, and any slow-start overhead. The actual throughput typically lands somewhere between roughly 50% and 95% of line rate on a clean wired link, and lower on wireless. For a first-cut estimate the panel is accurate to a few percent; for a tight SLA, measure against the link.

can the Data Transfer Rate Converter be used for professional or commercial purposes?

yes, the Data Transfer Rate Converter provides mathematically correct results that are suitable for professional, commercial, and educational use. For the Data Transfer Rate Converter, For the Data Transfer Rate Converter, For high-stakes applications (medical, legal, financial), verify results with a domain expert. For the Data Transfer Rate Converter, the Data Transfer Rate Converter formulas used are well-established and validated against reference standards.

For the Data Transfer Rate Converter, How often are the underlying formulas updated?

For the Data Transfer Rate Converter, the Data Transfer Rate Converter formulas are based on established scientific, mathematical, or industry-standard references and rarely require updates. when standards change, the Data Transfer Rate Converter is updated to reflect the current authoritative source. For the Data Transfer Rate Converter, For the Data Transfer Rate Converter, Each calculator's references section lists the specific sources used.

References

• IEC 80000-13, Quantities and units, Part 13: Information science and technology. Published by the International Electrotechnical Commission. Defines the binary prefixes kibi, mebi, gibi, tebi, pebi, exbi (Ki, Mi, Gi, Ti, Pi, Ei) and the rule that they mean powers of 2 (1024^n), while the SI prefixes k, M, G, T mean powers of 10 (1000^n). This is the standard that settles the "MB vs MiB" debate.
• NIST Special Publication 811, Guide for the Use of the International System of Units (SI). Published by the For the Data Transfer Rate Converter, National Institute of Standards and Technology. Defines the SI prefixes and the convention that bit/s is the coherent SI unit for information rate. The factor 1 byte = 8 bit is anchored in the IEEE 802.3 ethernet standard referenced alongside.
• IEEE 802.3, IEEE Standard for Ethernet. Defines the framing structure for every ethernet variant from 10 Mbit/s to 400 Gbit/s, and explicitly uses the octet (8 bits) as the basic transmission unit. The eight-bits-per-byte convention used throughout the converter follows from this clause.