Physical Layer
Open in the Padhaiko app2.1 Network monitoring: Bandwidth, throughput, delay, round-trip-time
The physical layer is tasked with transmitting raw, unstructured bitstreams across physical transmission media. Evaluating network performance requires quantifiable physical and operational metrics:
Bandwidth, Throughput, and Goodput
- Bandwidth (): In analog systems, bandwidth is the width of the frequency spectrum (measured in Hertz, ) that a transmission medium can pass without excessive signal attenuation. In digital networking, bandwidth denotes the theoretical maximum data carrying capacity of a link, measured in bits per second (, , , ).
- Throughput (): The actual rate at which useful data is successfully transferred across a communication channel under real-world conditions (accounting for protocol overhead, queuing, collisions, and retransmissions).
- Goodput: The rate of useful application-level data delivered to the destination process, excluding all protocol headers, retransmissions, and framing control bytes.
Components of Packet Delay (Latency)
The total one-way end-to-end delay () experienced by a packet traversing a communication path from source to destination across intermediate routers is the cumulative sum of four distinct delay components:
- Processing Delay (): The time required by a router or host CPU to inspect the packet header, verify the checksum or Frame Check Sequence (FCS), determine the outbound interface via routing table lookup, and decrement the Time-to-Live (TTL). Typically microseconds () in high-speed hardware ASICs.
- Queuing Delay (): The time a packet spends waiting in the router's outbound buffer queue before being scheduled for transmission. Queuing delay depends dynamically on the network traffic arrival intensity:
where is packet size in bits, is average packet arrival rate (), and is link transmission rate (). As , queuing delay increases asymptotically towards infinity.
- Transmission Delay (): The time required to push all bits of a packet onto the physical transmission link:
where is the packet length in bits and is the transmission speed of the link in bits per second. Transmission delay depends strictly on the packet length and link bit rate, completely independent of distance.
- Propagation Delay (): The time required for a physical signal (electromagnetic wave or optical pulse) to propagate across the physical length of the transmission medium from sender to receiver:
where is the physical distance of the link in meters and is the propagation speed of the signal in the medium (typically in copper and silica glass optical fiber, or in free-space radio/vacuum).
Round-Trip Time (RTT) and Bandwidth-Delay Product (BDP)
- Round-Trip Time (RTT): The total elapsed time from when a sender starts transmitting a packet until an acknowledgement (ACK) frame is received back from the destination. In a symmetric link neglecting processing delays, .
- Bandwidth-Delay Product (BDP): Defines the maximum volume of unacknowledged data bits that can be "in flight" inside the physical communication pipe at any given instant:
The BDP dictates the minimum buffer size and TCP receive window () necessary to achieve link utilization.
2.2 Multiplexing - FDM, TDM, WDM
Multiplexing
Multiplexing is the set of techniques that allows the simultaneous transmission of multiple independent information signals across a single shared high-capacity physical data link. A multiplexer (MUX) combines the incoming channels into a composite stream, and a demultiplexer (DEMUX) separates them at the receiving end.
| Aspect | Frequency Division Multiplexing (FDM) | Time Division Multiplexing (TDM) |
|---|---|---|
| Signal Domain | Analog signals (continuous frequency division) | Digital signals (discrete time slot allocation) |
| Channel Sharing Mechanism | Total channel bandwidth is divided into non-overlapping frequency sub-bands | Entire channel bandwidth is allocated to one channel at a time for a slice of time |
| Separation Buffers | Requires unused guard bands to avoid adjacent channel crosstalk | Requires guard times or frame synchronization bits to align slots |
| Hardware Complexity | Requires analog bandpass filters and distinct RF modulators per channel | Requires digital shift registers, buffers, and precise clock synchronization |
| Typical Applications | AM/FM radio, Cable TV broadcasting, ADSL broadband | T1/E1 digital telephony carriers, SONET/SDH, cellular GSM |
1. Frequency Division Multiplexing (FDM)
In FDM, each input signal modulates a distinct carrier frequency. The modulated signals are combined into a single wideband composite signal. To prevent overlapping and crosstalk between adjacent channels, unused frequency buffers called guard bands are inserted between carrier frequencies.
2. Time Division Multiplexing (TDM)
TDM operates on digital bitstreams. Time on the physical channel is divided into recurring frames, and each frame is further divided into fixed-duration time slots:
- Synchronous TDM (STDM): Each input line is pre-assigned a dedicated time slot in every recurring frame, whether the connected host has active data to transmit or not. If a terminal is idle, its assigned slot travels empty, wasting transmission bandwidth.
- Statistical (Asynchronous) TDM (ATDM): Slots are allocated dynamically on demand only to input lines that currently hold data in their buffers. Because slots are dynamically shared, each transmitted slot must carry an addressing header identifying the source channel, adding minor protocol overhead in exchange for vastly superior bandwidth utilization.
3. Wavelength Division Multiplexing (WDM)
Wavelength Division Multiplexing is the optical equivalent of FDM applied to fiber-optic cables:
- Multiple optical laser signals of differing wavelengths (, or colors of infrared light) are injected into a single optical fiber core simultaneously using an optical prism or diffraction grating.
- Coarse WDM (CWDM): Uses wide channel spacing (), supporting up to channels over short campus links without active laser cooling.
- Dense WDM (DWDM): Uses extremely narrow channel spacing ( or conforming to the ITU-T / grid), multiplexing over independent channels onto a single fiber pair, achieving aggregate throughputs exceeding .
2.3 Switching - circuit switching, packet switching, datagram and virtual circuit switching
When interconnecting multiple communicating hosts across a wide area, establishing dedicated direct physical links between every pair of nodes is economically impossible. Networks therefore employ switched networks where intermediate nodes (switches) redirect traffic along shared transmission trunks.
| Aspect | Circuit Switching | Packet Switching |
|---|---|---|
| Dedicated Path | Yes; a dedicated physical circuit is reserved end-to-end for the call duration | No; packets are routed dynamically over shared links using statistical multiplexing |
| Call Setup Phase | Mandatory end-to-end setup phase required before data transmission begins | No call setup for datagrams; virtual circuits require lightweight signaling setup |
| Data Flow | Continuous stream of bits; no store-and-forward delay at intermediate switches | Discrete packets; intermediate nodes store and forward packets |
| Bandwidth Reservation | Fixed, guaranteed bandwidth reserved throughout the entire call duration | Dynamic; bandwidth is consumed on demand; unused bandwidth is shared |
| Queuing Delay | Zero queuing delay after connection establishment | Variable queuing delay and risk of buffer overflow / packet drop during congestion |
| Fault Recovery | If an intermediate switch or link fails, the entire circuit disconnects | Packets can be routed around failed intermediate links dynamically |
| Primary Examples | Traditional Public Switched Telephone Network (PSTN), ISDN | The global Internet (IP), Ethernet, MPLS, ATM |
Packet Switching: Datagram vs. Virtual Circuit Approach
Packet-switched networks operate under two fundamentally contrasting architectural paradigms:
- Datagram Switching (Connectionless Packet Switching):
- Each individual packet (datagram) is treated as an independent self-contained transmission entity.
- Every datagram contains the complete destination logical (IP) address.
- Intermediate routers consult their internal routing tables independently for each packet; consequently, different packets belonging to the very same application session may follow completely different physical paths through the network.
- Packets may arrive at the destination out of order, duplicated, or delayed, requiring higher-layer protocols (TCP) to reorder and reassemble them.
- Advantage: Highly resilient; robust against switch or link failures.
- Virtual Circuit (VC) Switching (Connection-Oriented Packet Switching):
- A logical path called a Virtual Circuit is established between sender and receiver before any data packets are transmitted (using signaling protocols).
- All packets belonging to the connection follow the exact same predetermined physical path.
- Rather than carrying complete or destination addresses, each packet header carries only a compact local identifier: a Virtual Circuit Identifier (VCI) or Label.
- Intermediate switches maintain an internal VC translation table mapping
(incoming interface, incoming VCI) -> (outgoing interface, outgoing VCI). - Ensures guaranteed in-order packet delivery and rapid hardware label swapping at intermediate nodes.
- Prominent Examples: Asynchronous Transfer Mode (ATM), Multiprotocol Label Switching (MPLS), Frame Relay.
2.4 Transmission media - Guided (twisted pair, coaxial, optical fiber) and unguided media (radio waves, microwaves, infrared)
Transmission media represent the physical pathway that conveys electromagnetic or optical signals between transmitter and receiver. They are categorized into guided (bounded/wired) and unguided (unbounded/wireless) media:
Guided Transmission Media
- Twisted Pair Cable:
- Consists of two insulated copper conductor wires twisted around each other in a continuous regular helical pattern.
- Why Twisting? When external electromagnetic interference (EMI) or radio frequency interference (RFI) radiates across the cable, twisting ensures that both wires are equally exposed to the noise source. By utilizing differential signaling (the receiver measures ), equal noise induced on both conductors cancels out completely at the receiver.
- UTP (Unshielded Twisted Pair): Highly flexible, inexpensive, standard for office Ethernet cabling.
- STP (Shielded Twisted Pair): Encases pairs in metallic foil/braided shielding to block heavy industrial electromagnetic noise.
- Categories: Cat 5e (, supports up to ), Cat 6 (, supports up to ), Cat 6a (, supports up to ).
- Coaxial Cable:
- Consists of a central solid copper core conductor surrounded by a flexible dielectric insulator, encased in a cylindrical braided metallic mesh conductor, wrapped in an outer plastic protective sheath.
- The metallic braid provides excellent shielding against external EMI and prevents internal signal radiation.
- Varieties: baseband (used in legacy 10BASE5 Thicknet and 10BASE2 Thinnet Ethernet) and broadband (used for Cable Television and DOCSIS cable modems).
- Optical Fiber:
- Transmits information as modulated pulses of light through a central ultra-pure cylindrical core of silica glass () or plastic, surrounded by a glass cladding of lower refractive index (), protected by a polymer buffer coating and Kevlar strength members.
- Operates on the physical principle of Total Internal Reflection (TIR): when light strikes the core-cladding boundary at an angle of incidence greater than the critical angle , the light pulse is reflected back into the core with negligible refraction loss.
- Single-Mode Fiber (SMF): Has an extremely narrow core diameter () that allows only a single spatial mode of light to propagate along the central axis. Eliminates modal dispersion entirely; powered by semiconductor laser diodes ( or ); supports transmission distances over without repeaters.
- Multi-Mode Fiber (MMF): Has a wider core diameter ( or ) allowing light rays to bounce along multiple propagation angles (modes). Suffers from modal dispersion (different modes arrive at slightly different times, broadening the pulse); powered by inexpensive LEDs or VCSELs (); limited to short campus/data-center runs ().
| Medium | Typical Bandwidth | Maximum Segment Distance | EMI / Noise Immunity | Cost per Meter |
|---|---|---|---|---|
| UTP (Cat 6a) | Up to | Moderate (cancels via differential twisting) | Very Low | |
| Coaxial Cable | Up to | (Thin) / (Thick) | High (braided copper shielding) | Moderate |
| MMF Optical Fiber | to | Immune to electromagnetic interference | Moderate to High | |
| SMF Optical Fiber | to | Immune to electromagnetic interference | High (due to transceivers) |
Unguided (Wireless) Transmission Media
- Radio Waves ( to ):
- Omnidirectional propagation (radiates in all directions from an isotropic antenna).
- Easily penetrates walls, buildings, and obstacles, making it ideal for indoor/outdoor mobile communication (FM radio, cellular networks, paging).
- Susceptible to multipath interference and government spectrum licensing.
- Terrestrial and Satellite Microwaves ( to ):
- Highly directional line-of-sight propagation requiring focused parabolic dish antennas aligned accurately toward one another.
- Does not penetrate solid obstacles well; subject to atmospheric absorption and rain fade above .
- Deployed for high-capacity telephone carrier backhauls and geostationary/LEO satellite communications.
- Infrared ( to ):
- Extremely short-range directional line-of-sight communication.
- Cannot penetrate solid walls, providing superior physical layer security against eavesdropping outside a conference room (e.g., TV remote controls, IrDA interfaces).
2.5 Ethernet cable standards (twisted pair and optical fiber)
The IEEE 802.3 working group standardized physical layer Ethernet implementations utilizing a standardized nomenclature:
| Ethernet Standard | Formal Name | Transmission Medium | Max Segment Length | Encoding / Modulation Scheme |
|---|---|---|---|---|
| 10BASE-T | IEEE 802.3i | 2 pairs Cat 3/5 UTP | Manchester () | |
| 100BASE-TX | IEEE 802.3u | 2 pairs Cat 5 UTP | 4B/5B with MLT-3 () | |
| 100BASE-FX | IEEE 802.3u | 2 strands MMF () | () | 4B/5B with NRZ-I |
| 1000BASE-T | IEEE 802.3ab | 4 pairs Cat 5e/6 UTP | 4D-PAM5 ( on 4 pairs) | |
| 1000BASE-SX | IEEE 802.3z | 2 strands MMF () | 8B/10B with NRZ laser | |
| 1000BASE-LX | IEEE 802.3z | 2 strands SMF () | () | 8B/10B with long-wave laser |
| 10GBASE-T | IEEE 802.3an | 4 pairs Cat 6a UTP | 128-DSQ PAM16 | |
| 10GBASE-SR | IEEE 802.3ae | 2 strands OM3/OM4 MMF | 64B/66B ( laser) | |
| 10GBASE-LR | IEEE 802.3ae | 2 strands SMF () | 64B/66B ( DFB laser) |
2.6 Data encoding: Manchester, NRZ-I, MLT3, 4B/5B
Digital line encoding converts binary bits (s and s) into discrete physical electrical voltage levels or optical pulses suitable for transmission across a physical link.
Fundamental Challenges in Line Encoding
- Clock Synchronization (Self-Clocking): The receiver must synchronize its sampling clock with the incoming signal. Long runs of identical bits ( or ) cause constant voltage levels with zero transitions, causing the receiver's clock to drift and miscount bits.
- DC Component (Baseline Wander): Constant positive or negative voltages accumulate a direct current (DC) bias. Because transformers, AC-coupled transceivers, and capacitive filters reject DC, DC accumulation distorts the signal baseline, leading to bit detection errors.
- Baud Rate vs. Bit Rate: Bit rate is the number of bits transmitted per second (); baud rate (modulation rate) is the number of signal voltage level transitions per second. Bandwidth requirements scale directly with baud rate.
Primary Line Encoding Schemes
- NRZ-L (Non-Return-to-Zero Level):
- Bit is mapped to a positive voltage (), and bit is mapped to a negative voltage () (or vice versa).
- Limitation: Long sequences of consecutive s or s produce zero transitions, causing complete clock drift and heavy DC bias.
- NRZ-I (Non-Return-to-Zero Invert on Ones):
- A transition in signal level (high-to-low or low-to-high) occurs at the beginning of the bit boundary whenever bit is transmitted.
- No transition occurs when bit is transmitted.
- Advantage: Solves the polarity inversion problem (differential detection).
- Limitation: A long sequence of zeros () generates zero transitions, resulting in loss of clock synchronization.
- Manchester Encoding (IEEE 802.3 Standard):
- Every bit interval features a mandatory voltage transition precisely at the center of the bit cell:
- Bit : High-to-Low transition at mid-bit.
- Bit : Low-to-High transition at mid-bit.
- Advantage: Guarantees at least one transition per bit, providing self-clocking and zero DC bias.
- Disadvantage: Requires a baud rate of twice the bit rate (), achieving only bandwidth efficiency (e.g., at requires a signal bandwidth).
- Every bit interval features a mandatory voltage transition precisely at the center of the bit cell:
- Differential Manchester (IEEE 802.5 Token Ring):
- Always features a transition at the center of the bit cell for clock synchronization.
- Bit identity is encoded at the start of the bit interval:
- Bit : Features an explicit transition at the start of the bit interval.
- Bit : Features no transition at the start of the bit interval.
- Immune to accidental wire polarity reversal.
- MLT-3 (Multi-Line Transmission 3-Level):
- Utilizes three distinct signal voltage levels: , , and .
- Rules:
- If the next bit is , the signal level does not change.
- If the next bit is , the signal level transitions to the next state in the cyclic sequence:
- A complete cycle requires four transitions (four -bits), meaning the fundamental maximum signal frequency is only one-fourth of the bit rate ().
- Drastically reduces EMI and required copper bandwidth; used in Fast Ethernet (100BASE-TX).
- 4B/5B Block Coding:
- To prevent the clock drift of NRZ-I under long strings of zeros, 4B/5B block coding translates every data nibble into a pre-assigned symbol code word.
- The possible codes are chosen such that no code word contains more than one leading zero and no code word contains more than two trailing zeros.
- When transmitted back-to-back using NRZ-I, there can never be more than three consecutive zeros on the wire, ensuring regular signal transitions for clock recovery.
- Overhead: Transmission requires to send of data, yielding an efficiency of (only overhead, far superior to Manchester's penalty).
- In 100BASE-TX: user data is expanded by 4B/5B to , which is then modulated via MLT-3 over Cat 5 UTP.
Worked Out Examples
2.1
A source host transmits a data packet of size to destination host across a point-to-point fiber optic link characterized by:
- Link transmission bandwidth ().
- Physical link length ().
- Propagation speed of light in silica fiber .
- Intermediate routing processing delay .
- Average queuing delay .
Calculate:
- The transmission delay in milliseconds ().
- The propagation delay in milliseconds ().
- The total one-way end-to-end latency in milliseconds.
- The link Round-Trip Time (), assuming the return acknowledgement packet has negligible transmission delay.
- The Bandwidth-Delay Product () of the link in bits and in bytes.
Evaluating Transmission Delay and Propagation Delay:
- Packet length in bits:
- Transmission delay:
- Propagation delay across fiber:
Calculating Total One-Way Delay, RTT, and Bandwidth-Delay Product:
- Total one-way latency:
- Round-Trip Time (RTT):
- Bandwidth-Delay Product:
- Transmission delay: .
- Propagation delay: .
- Total one-way latency: .
- Link RTT: .
- Bandwidth-Delay Product: ().
2.2
A standard North American T1 carrier system utilizes Synchronous Time Division Multiplexing (STDM) to multiplex independent digitized voice channels onto a single digital line. Each voice channel is bandlimited to , sampled according to the Nyquist criterion at , and encoded using Pulse Code Modulation (PCM). In each recurring T1 frame, one 8-bit sample from each of the channels is interleaved sequentially, followed by additional synchronization framing bit.
Calculate:
- The total number of bits contained in a single T1 frame.
- The gross transmission bit rate of the T1 system in and .
- The net user data rate (payload capacity) in .
- The framing synchronization overhead as a percentage of total transmission capacity.
Calculating Frame Bit Structure and Gross Transmission Bit Rate:
- Bits contributed by the digitized voice channels:
- Total bits per T1 frame (including framing bit):
- Since each voice channel is sampled times per second, the multiplexer must transmit exactly to avoid buffer underflow or overflow:
- Gross transmission bit rate:
Evaluating Net Payload Rate and Synchronization Overhead:
- Net voice payload bit rate:
- Framing bit overhead rate:
- Overhead percentage:
- Net user data efficiency:
- Bits per T1 frame: .
- Gross transmission bit rate: ().
- Net payload bit rate: ().
- Framing overhead: (efficiency: ).
2.3
A telephone local loop transmission line has an analog channel bandwidth of ().
- Assuming an ideal noiseless transmission channel, calculate the maximum theoretical data rate achievable using distinct signaling voltage levels (Nyquist theorem).
- Now consider real-world thermal noise where the channel has a Signal-to-Noise Ratio of . Calculate the absolute theoretical upper bound on channel capacity according to Shannon's theorem.
Part 1 — Maximum Bit Rate on a Noiseless Channel (Nyquist Capacity): Nyquist's theorem states that for a channel with bandwidth transmitting a signal with discrete voltage levels:
Given and :
Part 2 — Maximum Channel Capacity Under Thermal Noise (Shannon Capacity): Shannon's capacity formula establishes the fundamental physical limit of error-free information transfer in the presence of Gaussian noise:
First convert into linear power ratio :
Evaluate channel capacity:
Using the base-conversion formula :
- Maximum Nyquist bit rate on noiseless channel: ().
- Maximum Shannon capacity at SNR: ().
Practice Exercises
Practice Exercises — Physical Layer & Transmission
Compare Circuit Switching, Datagram Packet Switching, and Virtual Circuit Packet Switching with respect to:
- Dedicated path setup requirement.
- Overhead per packet/frame.
- Intermediate switch routing complexity.
- Handling of intermediate switch or link failures.
Comparison Matrix of Switching Techniques:
| Dimension | Circuit Switching | Datagram Packet Switching | Virtual Circuit (VC) Switching |
|---|---|---|---|
| Path Setup | Mandatory end-to-end circuit reservation before data transfer | None (completely connectionless; immediate transmission) | Mandatory connection setup phase to establish VC table mappings |
| Header Overhead | Zero header overhead during data phase (continuous bit stream) | High (every datagram carries full source and destination IP addresses) | Low (packet carries only a short local identifier: VCI or Label) |
| Routing Decisions | Determined once during call setup; switches only forward along fixed circuit | Every router evaluates full routing table independently for every packet | Fast label swap (in-port, in-VCI) -> (out-port, out-VCI) in hardware |
| Failure Handling | Catastrophic; if any intermediate switch fails, call drops completely | Highly resilient; routers dynamically reroute subsequent packets around failures | Fragile; if a link fails, all VCs traversing that link must be re-established |
| Bandwidth Guarantee | Guaranteed, fixed allocated bandwidth (idle periods waste capacity) | Best-effort; statistical multiplexing shares capacity dynamically | Predictable QoS; bandwidth can be reserved along the virtual circuit |
Circuit switching requires mandatory physical path reservation with zero packet headers but cannot recover dynamically from link failure. Datagram switching requires no setup and recovers dynamically around failures, but incurs high header overhead and variable queuing delays. Virtual circuit switching combines connection setup with fast hardware label-switching and lower header overhead, but requires circuit re-establishment if an intermediate link fails.
A source host transmits a packet of size to a destination host across a network path traversing store-and-forward packet switches (forming identical successive transmission links). Each link has:
- Transmission bandwidth ().
- Physical length ().
- Propagation velocity .
Assume each intermediate switch begins transmitting a packet only after receiving it completely. Neglect processing and queuing delays.
Calculate:
- The transmission delay across a single link.
- The propagation delay across a single link.
- The total end-to-end latency to deliver the packet from host to host .
Evaluating Single-Link Delays:
- Packet length in bits:
- Single-link transmission delay:
- Single-link propagation delay:
Calculating Total Store-and-Forward End-to-End Latency: There are identical sequential links connecting source through store-and-forward switches to destination :
- Host transmits packet onto Link 1: takes .
- Switch 1 receives, then transmits packet onto Link 2: takes .
- Switch 2 receives, then transmits packet onto Link 3: takes .
- Switch 3 receives, then transmits packet onto Link 4: takes .
Total end-to-end delay:
- Transmission delay per link: .
- Propagation delay per link: .
- Total end-to-end store-and-forward latency across all 4 links: .
Compare Twisted Pair, Coaxial Cable, and Optical Fiber as guided transmission media with respect to:
- Physical construction and noise suppression mechanism.
- Signal attenuation and maximum unamplified segment distance.
- Susceptibility to electromagnetic interference (EMI) and wiretapping.
- Relative material and installation cost.
Comparative Breakdown of Guided Media:
- Physical Construction and Noise Suppression:
- Twisted Pair: Two insulated copper wires twisted helically. Cancels noise via differential signaling (equal external EMI induced on both wires is subtracted at receiver).
- Coaxial Cable: Concentric solid copper core, dielectric insulator, and woven copper braided outer shield that blocks external electromagnetic radiation.
- Optical Fiber: Ultra-pure silica glass core () surrounded by glass cladding (). Confines light via total internal reflection.
- Attenuation and Distance:
- Twisted Pair: High attenuation at higher frequencies; limited to in standard Ethernet.
- Coaxial Cable: Moderate attenuation; reaches (Thinnet) or (Thicknet).
- Optical Fiber: Extremely low attenuation ( at ); reaches without active amplification in single-mode fiber.
- EMI Immunity and Security:
- Twisted pair has moderate EMI resistance and radiates detectable RF signals (easy to wiretap). Coaxial has high EMI resistance. Optical fiber is completely immune to EMI/RFI/lightning and does not radiate electromagnetic energy, making unauthorized physical wiretapping exceptionally difficult.
- Cost:
- Twisted pair is the cheapest to purchase and terminate; coaxial is moderate; optical fiber cable itself is inexpensive, but optical transceivers (lasers/photodiodes) and precision fusion splicing require high initial capital.
Twisted pair uses differential helical twists ( limit, low cost); coaxial uses a grounded outer copper braid (, moderate cost); optical fiber uses total internal reflection of light pulses in silica glass (, complete EMI immunity, highest security, higher transceiver cost).
Explain the operating principles of NRZ-I, Manchester, and MLT-3 line encoding schemes. Explain why Manchester encoding is rarely used in Gigabit Ethernet and describe how modern high-speed Ethernet overcomes the baud-rate bottleneck.
Operating Principles of the Three Encoding Schemes:
- NRZ-I (Non-Return-to-Zero Inverted): Signal transitions occur at the start of the bit boundary on binary ; no transition occurs on binary . Requires external block coding (such as 4B/5B) to prevent clock drift during consecutive zeros.
- Manchester Encoding: Always introduces a transition at the midpoint of each bit interval (High-to-Low for , Low-to-High for ). The transition provides continuous self-clocking and eliminates DC bias.
- MLT-3 (Multi-Line Transmission 3-Level): Cycles through three voltage levels (). Transitions to the next level only when a binary is transmitted; remains constant on binary .
Why Manchester is Unsuitable for Gigabit Ethernet and Modern Solutions:
- In Manchester encoding, every single bit requires at least one transition, and bit sequences like require two transitions per bit. Consequently, transmitting at would require a modulation rate of ( bandwidth), far exceeding the transmission capabilities of standard Cat 5e/Cat 6 copper cabling.
- Modern high-speed standards overcome this bottleneck by:
- Block Encoding (4B/5B, 8B/10B, 64B/66B): Adds controlled redundancy to guarantee transitions without doubling frequency.
- Multilevel Signaling (PAM-5, PAM-16): Encodes multiple bits into a single symbol (e.g., in 1000BASE-T, PAM-5 encodes , running across all wire pairs simultaneously at only ).
NRZ-I transitions on ; Manchester transitions at mid-bit for clocking; MLT-3 cycles through 3 voltage states on . Manchester requires the bandwidth (2 Baud/bit), making it unfeasible at gigabit speeds. Gigabit Ethernet overcomes this using multilevel pulse amplitude modulation (PAM-5) and transmitting concurrently over all 4 pairs at a modest .
Fast Ethernet (100BASE-TX) uses a two-stage encoding pipeline: 4B/5B block coding followed by MLT-3 line encoding over Category 5 UTP cable.
- Explain why 4B/5B block coding is performed prior to MLT-3 transmission.
- Determine the baud rate on the physical cable required to achieve a net data rate of .
- Calculate the fundamental maximum signal frequency of the MLT-3 waveform on the wire.
- Calculate the coding efficiency and protocol overhead introduced by 4B/5B.
Purpose of 4B/5B Block Coding Before MLT-3:
- MLT-3 transitions only when transmitting a binary . If the raw data stream contains a long string of consecutive binary s, MLT-3 outputs a constant flat voltage with zero transitions, causing the receiver's phase-locked loop (PLL) clock to lose synchronization.
- 4B/5B maps every nibble to a symbol chosen such that no symbol has more than one leading zero and no symbol has more than two trailing zeros. Thus, when symbols are transmitted sequentially, there are never more than three consecutive zeros, guaranteeing frequent -bits that force regular MLT-3 transitions for clock recovery.
Calculating Baud Rate, Frequency, and Efficiency:
- Expanding through 4B/5B:
- Since MLT-3 generates at most one transition per encoded bit, the symbol rate (baud rate) is:
- A complete MLT-3 cycle () requires four consecutive -bits. Thus, the fundamental maximum electrical frequency is:
This easily fits within the bandwidth rating of Category 5 UTP cabling.
- Coding Efficiency:
- 4B/5B guarantees no more than 3 consecutive zeros, preventing receiver clock drift in MLT-3.
- Baud rate on the cable: .
- Maximum fundamental signal frequency: .
- Coding efficiency: ( overhead).