Fast Ethernet (IEEE 802.3) Physical Layer: Structure and Implementation Types

Fast Ethernet is the IEEE 802.3 family’s $100\ \text{Mb/s}$ Ethernet variant, standardized in IEEE 802.3u (“Fast Ethernet”) and specified across multiple physical implementations (often called “100BASE” types). At a high level, it keeps the same MAC concepts as classic Ethernet (frame format, MAC addressing, and CSMA/CD behavior where applicable) while changing the Physical Layer to support $100\ \text{Mb/s}$ signaling and media-specific variants such as keyword100BASE-TX, 100BASE-FX, and 100BASE-T4.

Mermaid overview of where Fast Ethernet fits:

[CalloutBlock]
type: "info"
title: "Key idea"
content: "Fast Ethernet mainly changes the Physical Layer signaling and media; the MAC is comparatively consistent with Ethernet, but the physical interface and signaling rates differ."

Footnotes

1. IEEE 802.3u (Fast Ethernet) overview and 100BASE variants: https://en.wikipedia.org/wiki/Fast_Ethernet - Covers 100BASE-TX/FX/T4 context and Fast Ethernet media variants. ↩

1) Physical Layer structure for Fast Ethernet

Fast Ethernet’s Physical Layer is structured around: (1) the physical signaling method (line code, clocking/encoding), (2) the media attachment unit for the chosen cable/fiber type (copper vs fiber), and (3) channel constraints (e.g., pair usage, bandwidth, reach). Different “100BASE-*” types share the same overall goal—reaching $100\ \text{Mb/s}$—but do so with different electrical/optical methods and different cabling requirements.2

Physical-layer building blocks (conceptual)

1. Transmitter (Tx): converts MAC-bitstream to line-coded symbols suitable for the medium (copper electrical or fiber optical).
2. Channel: the actual physical medium (UTP pairs or fiber).
3. Receiver (Rx): recovers the symbol clock (or uses recovered/embedded timing depending on signaling), detects symbols, and outputs bits to the MAC-side interface.
4. Management/Interface details: includes connector/cabling assumptions and, in many implementations, auto-negotiation behavior is handled by higher-level PHY management in later evolutions (Fast Ethernet itself is tied to IEEE 802.3u PHY definitions).

keywordLine coding is the defining aspect here: it determines how bits map to voltage/light transitions, which in turn constrains frequency spectrum and allowable cable distance.

Footnotes

1. IEEE 802.3u (Fast Ethernet) overview and 100BASE variants: https://en.wikipedia.org/wiki/Fast_Ethernet - Covers 100BASE-TX/FX/T4 context and Fast Ethernet media variants. ↩ ↩²

2. 100BASE-TX overview (UTP copper, 2-pair focus): https://en.wikipedia.org/wiki/Fast_Ethernet#100BASE-TX - Summarizes 100BASE-TX medium and characteristics. ↩

1.1) Media implementations (“100BASE” types)

Fast Ethernet defines several physical implementations. The most common are:

• keyword100BASE-TX (typically two pairs of Category 5 UTP)
• keyword100BASE-FX (typically multimode fiber)
• keyword100BASE-T4 (uses four pairs of older-grade UTP such as Category 3, depending on implementation)

These “types” differ primarily in pair usage and signaling method.

2) Detailed structure: Physical-layer signaling and encoding (what the PHY actually does)

Fast Ethernet PHY operation can be described as a pipeline:

1. MAC to PHY mapping: MAC outputs a serial/parallel data stream (implementation-specific interface).
2. Encoding / symbol mapping: PHY encodes bits into a line code suitable for the medium.
3. Clocking: PHY must maintain timing so receivers can recover the stream reliably at $100\ \text{Mb/s}$.
4. Analog transmission: copper drive levels or fiber optical modulation produce the physical waveform/light pulses.
5. Detection and decoding: receiver recovers timing and converts from line symbols back to bitstream.

2.1) Why encoding matters

At $100\ \text{Mb/s}$, raw NRZ-like transmission can suffer from bandwidth limits, poor clock recovery, and intersymbol interference. Therefore, 100BASE PHYs use line coding/encoding appropriate for reliable recovery over the specified channel and distance.

[CalloutBlock]
type: "tip"
title: "Mental model for exam-style questions"
content: "When asked about Fast Ethernet physical structure, identify (a) medium type (copper/fiber), (b) which 100BASE variant, and (c) the encoding/line coding approach that makes $100\ \text{Mb/s}$ reliable on that medium."

3) Implementation types at the physical layer

3.1) 100BASE-TX (Fast Ethernet over UTP)

keyword100BASE-TX is the most widespread Fast Ethernet copper variant. It uses:

• 2-pair UTP (commonly Category 5-class cabling in deployment),
• an electrical PHY with specific line signaling/encoding compatible with $100\ \text{Mb/s}$,
• a typical duplex operation (full-duplex reduces/avoids collision-domain behavior in switched networks).

Conceptual interface view:

Practical implication: 100BASE-TX was designed to work over structured copper cabling standards that meet the required bandwidth for $100\ \text{Mb/s}$.

---

3.2) 100BASE-FX (Fast Ethernet over fiber)

keyword100BASE-FX replaces copper electrical signaling with optical signaling over fiber.

Key physical-layer structure differences:

• Optical transmitter (LED/laser depending on implementation) modulates light.
• Optical receiver detects light levels and converts back to electrical signals.
• Fiber provides different channel impairments (optical loss, dispersion) than copper (attenuation at high frequency, crosstalk).

Conceptual model:

Practical implication: fiber variants typically enable longer reach and provide immunity to some electromagnetic interference issues inherent to copper.

3.3) 100BASE-T4 (Fast Ethernet over older UTP using 4 pairs)

keyword100BASE-T4 was designed to support $100\ \text{Mb/s}$ Ethernet over cabling that might not have been high-quality enough for 100BASE-TX.

Physical-layer structure highlights:

• Uses 4 pairs of copper instead of 2.
• Splits and/or combines signaling across multiple pairs to meet performance over the channel.
• Requires more complex PHY processing than 2-pair designs.

Conceptual view:

Practical implication: 100BASE-T4 is largely a “migration/fallback” approach to avoid immediately replacing cabling in some environments.

4) Physical layer and collision-domain consequences (hub vs switch)

While the Physical Layer is distinct from MAC, the choice of medium and switching/hubbing affects how collisions are handled in practice.

In classic Ethernet with hubs, multiple stations share a collision domain; collisions affect throughput and require CSMA/CD. In modern networks with switches, each link is its own collision domain, so CSMA/CD effects are reduced or eliminated in practice (depending on duplex and configuration). Fast Ethernet’s PHY still interacts with MAC timing such as minimum packet size constraints and slot-time behavior at $100\ \text{Mb/s}$, which is why Ethernet timing rules matter when you discuss “physical-layer + Ethernet behavior” together.

[CalloutBlock]
type: "warning"
title: "Common misconception"
content: "The PHY type (TX vs FX vs T4) does not by itself “turn off” collisions; collision behavior depends on hub vs switch topology and duplex mode, even though the PHY defines the signaling on the link."

Footnotes

1. Ethernet collision domains / CSMA/CD and effect of hubs vs switches: https://en.wikipedia.org/wiki/Carrier-sense_multiple_access_with_collision_detection - Explains collision behavior and CSMA/CD fundamentals that tie to physical/MAC interaction in shared media. ↩

5) Quick comparison table (physical-layer focus)