Once upon a time, Ethernet was the plucky underdog of networking, running on thick coaxial cables with a MAC (Media Access Control) layer that was simple, collision-prone, and designed for speeds that would make modern engineers laugh. Fast forward to today, and Ethernet MAC has undergone a complete transformation, particularly when it comes to fibre-based implementations. The Ethernet of today is nothing like its ancestor—it's faster, more efficient, and in many ways, closer in spirit to technologies like FDDI (Fibre Distributed Data Interface) than to its own roots.
The Early Days | Simplicity & Collisions
Back in the day, Ethernet relied on CSMA/CD (Carrier Sense Multiple Access with Collision Detection). This meant every device on a shared bus would listen before transmitting, but if two devices transmitted at the same time—boom!—a collision occurred, and both devices had to back off and try again.
The MAC layer was relatively simple, using a flat addressing scheme and managing access to the network in a straightforward but inefficient manner. Speeds were low (10 Mbps was cutting-edge), and the biggest concern was just making sure packets didn’t bump into each other too often.
The Need for Change | Enter Switching & Higher Speeds
As Ethernet speeds increased (100 Mbps, then 1 Gbps, then 10 Gbps and beyond), the old MAC approach couldn’t keep up. Collisions became impractical, and the idea of a shared medium gave way to switching, where each device got a dedicated path to the network. This eliminated the need for CSMA/CD and made the MAC layer more about forwarding efficiency than collision management.
Fibre Ethernet, in particular, forced even more drastic changes. Unlike copper-based Ethernet, fibre links don’t have the same physical limitations. High-speed fibre Ethernet operates more like traditional circuit-switched networking, where deterministic access and high throughput are prioritised over legacy contention-based models.
High-Speed MAC | More FDDI Than Ethernet
At extreme speeds (25 Gbps, 40 Gbps, 100 Gbps, and beyond), modern Ethernet MAC functions in a way that would make early Ethernet engineers scratch their heads. Here’s how it resembles FDDI more than old-school Ethernet:
Deterministic Transmission: Traditional Ethernet was based on contention, but high-speed Ethernet has moved towards scheduled transmission mechanisms, similar to FDDI's token-passing concept.
Error Handling: Legacy Ethernet relied on CSMA/CD to detect and retransmit after collisions. Modern MAC layers use Forward Error Correction (FEC) and sophisticated buffering to avoid retransmissions.
Frame Structure Enhancements: While Ethernet frames remain similar in structure, high-speed MAC incorporates additional mechanisms for Quality of Service (QoS), frame pre-emption, and prioritisation, akin to what FDDI did to ensure real-time traffic.
Full Duplex by Default: Early Ethernet was half-duplex, meaning devices had to take turns talking. High-speed Ethernet, especially over fibre, is always full-duplex, eliminating the inefficiencies of the old model.
Why This Matters
The transformation of the Ethernet MAC is not just a technical curiosity—it’s a fundamental shift that has enabled modern networking. The shift away from contention-based access and towards a structured, high-performance model means that Ethernet can now support cloud data centres, AI workloads, and ultra-low latency financial transactions. Without these advancements, we’d still be stuck troubleshooting collisions instead of pushing multi-terabit traffic across continents.
In many ways, today’s Ethernet MAC is what FDDI tried to be—fast, reliable, and built for fibre. The difference? Ethernet won the battle of networking standards, while FDDI faded into history. But if you look under the hood of your 400 Gbps Ethernet MAC, you might just find an FDDI engineer smiling back at you. 😆💡🚀
