Why a 1 Gbps Ethernet Link Rarely Hits 125 MB/s and How Jumbo Frames Help
This article explains the difference between decimal and binary data units, how Ethernet frame overhead and protocol headers reduce the theoretical 125 MB/s throughput of a 1 Gbps link, and how increasing the MTU with jumbo frames can improve efficiency.
When transferring data over a network, the unit Mbps is commonly used, and Gigabit Ethernet (1 GbE) cards—often called “gigabit network cards”—are popular.
People also use GB or MB to describe disk size (Gigabyte or Megabyte), but when discussing file size or disk space they often use GiB or MiB, which follow binary prefixes.
The distinction between GB and GiB stems from decimal (10³) versus binary (2¹⁰) counting systems. The short‑scale system (thousands, millions, billions) is standard in most of Europe and the United States, while the long‑scale uses powers of a million. To avoid confusion, the International Organization for Standardization defined standardized prefixes based on powers of 10³ (kilo, mega, giga, tera, etc.). In binary, the same prefixes are altered to Gibi, Tebi, etc., giving us MiB for mebibyte.
For a 1000 Mbps network, the theoretical conversion to megabytes is 125 MB/s because 1 Gbps equals 1 000 000 000 bits per second, and 8 bits make a byte.
In practice, however, data must be encapsulated into Ethernet frames, which adds overhead. An Ethernet frame consists of a 7‑byte preamble, header fields, up to 1500 bytes of payload, and a 4‑byte CRC, totaling 1538 bytes. This overhead reduces the effective payload efficiency to 1500 / (1500 + 38) ≈ 97.53 % (without VLAN). Adding a 4‑byte VLAN tag lowers it to 97.28 %.
Consequently, a gigabit Ethernet link typically achieves about 975.3 Mbps (or 972.8 Mbps with VLAN) of usable data rate, not the full 1000 Mbps.
The SMB protocol used for file sharing adds further headers, so the actual file‑transfer speed is lower than the 121.9 MB/s derived from 975.3 Mbps.
To approach the theoretical limit, one can increase the MTU size—using jumbo frames up to 9000 bytes. The efficiency then becomes 9000 / (9000 + 38) ≈ 99.58 %, offering higher throughput for large data transfers.
However, small packets still consume a full frame, wasting bandwidth. The common solution is to separate traffic: use jumbo‑frame‑enabled switches for bulk data and standard‑MTU switches for small control packets, often separating storage networks (e.g., iSCSI, Fibre Channel) from regular business traffic.
Signed-in readers can open the original source through BestHub's protected redirect.
This article has been distilled and summarized from source material, then republished for learning and reference. If you believe it infringes your rights, please contactand we will review it promptly.
Open Source Linux
Focused on sharing Linux/Unix content, covering fundamentals, system development, network programming, automation/operations, cloud computing, and related professional knowledge.
How this landed with the community
Was this worth your time?
0 Comments
Thoughtful readers leave field notes, pushback, and hard-won operational detail here.
