How this instrument works
Gigabit-class networking is where most enterprise and carrier infrastructure now lives, so Gbps is the natural unit for describing it, while Mbps persists wherever a smaller circuit, an older standard, or a per-application allocation is being sized. The relationship between them is pure SI decimal scaling: giga is exactly one thousand mega, so 1000 Mbps equals exactly 1 Gbps with no rounding involved, the same convention that makes 1000 metres a kilometre.
Ethernet's own history traces this exact progression in hardware. 10BASE-T ran at 10 Mbps in the 1990s; Fast Ethernet pushed that to 100 Mbps; Gigabit Ethernet (1000BASE-T) reached 1000 Mbps, or 1 Gbps, becoming the standard wired port speed on most business and consumer equipment for years. Data-centre backbones and carrier links moved past that long ago: 10GbE, 40GbE and 100GbE are now common inside data centres, and long-haul fibre backbones run at multiples of 100 Gbps between major routing points — all of it expressed in Gbps because writing '100,000 Mbps' for a 100GbE link would be unreadable.
The conversion still matters at the boundary between those worlds. A cloud provider might bill egress bandwidth in Mbps for a small virtual machine but cap a dedicated interconnect at 10 Gbps; a network engineer sizing WAN capacity has to move between a branch office's 500 Mbps circuit and a data-centre's 10 Gbps core switch fabric in the same planning document, and getting that factor of 1000 wrong by even one step is the difference between a link that is undersized and one that is absurdly oversized.
- Enter the link or circuit speed in Megabits/second (Mbps) into that field; it opens at 500, a common business-grade WAN circuit size.
- Read the Gigabits/second (Gbps) line beneath it — recalculated on every keystroke, exact to three decimal places.
- Sizing capacity against a data-centre switch or backbone rated in Gbps? Convert each branch or circuit speed here first so every figure in your plan shares one unit.
- Going the other way from a Gbps spec sheet, multiply by 1000 to get Mbps — a 10 Gbps uplink is 10,000 Mbps.
- A negative entry is rejected here, since a data rate cannot fall below zero.
Worked example — a 500 Mbps branch circuit against a 10 Gbps core
A regional office is provisioned with a 500 Mbps dedicated internet circuit, a common size for a mid-sized business location. Enter 500 into Megabits/second (Mbps) and the Gigabits/second (Gbps) line returns 0.5 — half a gigabit, a figure that is easy to compare directly against a data-centre's core switching capacity once both are in the same unit.
If that office's traffic all lands on a data-centre uplink provisioned at 10 Gbps (10,000 Mbps), the branch circuit represents exactly 5% of that uplink's total capacity — 0.5 Gbps against 10 Gbps. A network planner sizing that core link across twenty similar branch offices would need roughly 10 Gbps of aggregate capacity just to match this one office's peak, before accounting for any other traffic sharing the same backbone.
Questions
Is 1000 Mbps exactly 1 Gbps, or is that a rounded approximation?
Exactly 1 Gbps. Networking prefixes follow the SI decimal convention throughout — kilo, mega, giga each a factor of exactly 1000 apart — so 1000 Mbps = 1 Gbps with zero rounding at the definitional level. Some Gigabit Ethernet hardware falls slightly short of that theoretical maximum in practice due to protocol overhead, but that is a real-world throughput limit, not an error in the unit conversion itself.
How did Ethernet speeds progress from Mbps to Gbps?
In roughly ten-times steps: 10BASE-T at 10 Mbps in the late 1980s and 1990s, Fast Ethernet at 100 Mbps through the mid-to-late 1990s, then Gigabit Ethernet (1000BASE-T) at 1000 Mbps — 1 Gbps — becoming the standard wired port speed on most equipment by the mid-2000s. From there, data centres moved to 10, 40 and 100 Gbps standards well before consumer equipment caught up, since backbone and switch-fabric demand grew faster than desktop networking needs.
Why do cloud providers sometimes bill bandwidth in Mbps and sometimes in Gbps?
It generally tracks the size of what is being described. A small virtual machine or a metered egress allowance is often capped or billed in Mbps because the numbers stay in a readable three- or four-digit range; a dedicated interconnect, a reserved network circuit, or a data-centre cross-connect is usually specified in Gbps because the underlying capacity is in the multiple-gigabit range, where an Mbps figure would run to five or six digits.
What Gbps speed is a 'gigabit internet' home plan?
Typically right around 1 Gbps (1000 Mbps) download, sometimes with a lower upload figure depending on the technology — cable, fibre or fixed wireless. Some newer fibre plans now advertise multi-gigabit tiers such as 2 Gbps or 5 Gbps. In every case the plan's Mbps or Gbps rating describes the maximum rate the provider's network delivers to the home, not the speed any single device or Wi-Fi connection inside the house will necessarily achieve.
What is the difference between 10GbE and a 10 Gbps internet plan?
The unit and rated speed are identical, but the context differs. 10GbE refers to the IEEE 802.3ae Ethernet standard used for switch uplinks, server network interfaces and data-centre fabric, running over fibre or short-reach copper. A '10 Gbps' consumer or business internet plan describes the rate a provider delivers over its access network, which may or may not use 10GbE hardware at either end — the two terms describe the same rate but different layers of infrastructure.
How much Mbps does a fully loaded 10 Gbps link actually carry?
10,000 Mbps at full theoretical capacity, since 10 Gbps × 1000 converts directly. In practice, protocol overhead, error correction and shared contention mean sustained real-world throughput usually runs somewhat below that nominal figure, the same way a 100 Mbps home connection rarely sustains a full 100 Mbps of usable file-transfer speed — the unit conversion is exact, but real network utilisation is a separate, lower number.