Mbps to MB/s Converter

1. The Fundamental Distinction: Bits (b) vs. Bytes (B)

The most pervasive point of confusion among internet subscribers, systems administrators, and junior network engineers stems from the typographical and mathematical distinction between the lowercase letter 'b' (bit) and the uppercase letter 'B' (Byte).

A bit (binary digit) is the atomic unit of digital communications. Physical transceivers—such as optical SFP+ lasers, DSL modems, DOCSIS cable modems, and 5G baseband chips—transmit data serially or in parallel as individual electromagnetic pulses, light flashes, or RF phase shifts. Because line hardware operates sequentially on bit streams, telecommunications standards organizations (IEEE, ITU-T, 3GPP) and Internet Service Providers (ISPs) universally specify connection bandwidth in Megabits per second (Mbps) or Gigabits per second (Gbps).

In contrast, a Byte is a composite structural unit consisting of exactly 8 bits ($1\text{ Byte} = 8\text{ bits}$). Operating systems (Microsoft Windows, Apple macOS, Linux), computer storage controllers, SSDs, hard drives, and application downloaders (Steam, BitTorrent, Google Chrome) quantify memory addresses and file sizes strictly in Megabytes (MB) or Gigabytes (GB).

Consequently, the fundamental rule of thumb for computing download capacity is the 8-to-1 divisor:

If a consumer purchases an advertised "100 Mbps" broadband plan, the maximum raw theoretical throughput is:

Under no circumstances will a 100 Mbps broadband connection yield a sustained download speed of 100 Megabytes per second.

2. The Base-10 (SI) vs. Base-2 (IEC) Discrepancy (MB/s vs. MiB/s)

Even after dividing advertised Mbps by 8, network users routinely observe that their file downloads plateau at a number approximately $5\%$ lower than expected. On a 100 Mbps connection, for example, a Windows Explorer file copy or Steam client often caps out at approximately $11.92\text{ MB/s}$. This behavior is not necessarily caused by network congestion, but rather by an unresolved conflict between decimal and binary unit standards:

• International System of Units (SI Metric / Telecom Standard): Uses powers of 10.
  $1\text{ kilobit (kb)} = 10^3\text{ bits} = 1,000\text{ bits}$
  $1\text{ Megabit (Mb)} = 10^6\text{ bits} = 1,000,000\text{ bits}$
  $1\text{ Megabyte (MB)} = 10^6\text{ Bytes} = 1,000,000\text{ Bytes}$
• International Electrotechnical Commission (IEC 80000-13 / Binary Standard): Uses powers of 2.
  $1\text{ Kibibyte (KiB)} = 2^{10}\text{ Bytes} = 1,024\text{ Bytes}$
  $1\text{ Mebibyte (MiB)} = 2^{20}\text{ Bytes} = 1,048,576\text{ Bytes}$
  $1\text{ Gibibyte (GiB)} = 2^{30}\text{ Bytes} = 1,073,741,824\text{ Bytes}$

3. Protocol Encapsulation Tax: The Missing 5% to 10% Goodput

In addition to the binary-to-decimal numbering difference, physical networks must consume bandwidth to transmit protocol control headers. Every user byte transferred over the internet must be wrapped inside Layer 2 Ethernet, Layer 3 IP, and Layer 4 TCP headers:

• Physical Layer 1 Framing: 7 Bytes Preamble + 1 Byte Start Frame Delimiter (SFD) + 12 Bytes Inter-Packet Gap (IPG) = 20 Bytes on the wire.
• Layer 2 Ethernet Frame: 14 Bytes MAC Header (Destination MAC, Source MAC, EtherType) + 4 Bytes Frame Check Sequence (FCS) = 18 Bytes.
• Layer 3 IP Packet: 20 Bytes for IPv4 (or 40 Bytes for IPv6).
• Layer 4 TCP Segment: 20 Bytes minimum header (source/dest ports, sequence numbers, flags, window size).

On a standard internet link with a Maximum Transmission Unit (MTU) of 1,500 Bytes, the maximum usable TCP payload—termed the Maximum Segment Size (MSS)—is:

However, the total physical wire footprint required to convey this 1,460-byte payload across Ethernet is:

Thus, the maximum theoretical protocol efficiency ($\eta$) of TCP/IPv4 over standard Ethernet is:

This $5.07\%$ loss is the unavoidable "protocol encapsulation tax". When combined with the $4.86\%$ binary shift, a mathematically perfect 100 Mbps line can deliver at most:

When factoring in additional real-world constraints—such as TLS 1.3 cryptographic record headers (5 to 16 Bytes per block), HTTP/2 frame overhead, Wi-Fi channel contention, and TCP slow-start window ramping—practical consumer download goodput typically ranges between $88\%$ and $92\%$ of the ISP's advertised line rate.