DATA COMMUNICATION (EC-603 (A)) - Unit 3 Short Notes

I. FUNDAMENTALS OF DATA COMMUNICATION

A. Data Communication System Components

• Source: Generates data to be transmitted (e.g., computer, sensor).

• Transmitter: Converts data into transmittable signals (e.g., modem, encoder).

• Medium: Physical path for signal propagation (guided/unguided).

• Receiver: Captures and reconstructs original data from signals.

• Destination: Intended endpoint of communication.

B. Characteristics of Data Communication

• Delivery: Data reaches correct destination.

• Accuracy: Error-free transmission (bit error rate).

• Timeliness: Data delivered within acceptable delay.

• Jitter: Variation in packet arrival time (critical for real-time apps).

C. Network Types and Classification

D. Transmission Modes

• Simplex: One-way communication (e.g., keyboard → CPU, TV broadcast).

• Half-duplex: Two-way but not simultaneous (e.g., walkie-talkie, CSMA/CD).

• Full-duplex: Simultaneous two-way (e.g., telephone, Ethernet switch).

E. Serial vs Parallel Transmission

• Serial: Bits sent sequentially over single channel. Used for long distances (USB, Ethernet). Requires fewer wires, lower cost.

• Parallel: Multiple bits sent simultaneously over multiple channels. Used for short distances (printer port, internal bus). Higher speed but prone to skew, crosstalk.

F. Transmission Media

Guided Media:

• Twisted Pair: UTP/STP, up to 1 Gbps, susceptible to EMI, cheap.

• Coaxial Cable: Thick/thin, 10–100 Mbps, better shielding than twisted pair.

• Fiber Optic: Single-mode (long-haul, high BW) / Multi-mode (short-haul). Low attenuation, high BW, immune to EMI.

Unguided Media:

• Radio: WiFi, cellular, omnidirectional, subject to interference.

• Microwave: Point-to-point, line-of-sight, high BW, weather-sensitive.

• Infrared: Short-range, line-of-sight, used in remote controls.

G. Transmission Impairments

• Attenuation: Signal strength loss with distance. Mitigation: Amplifiers/repeaters.

• Distortion: Signal shape change due to dispersion. Mitigation: Equalization.

• Noise:

  • Thermal: Random electron motion (white noise).

  • Intermodulation: Nonlinear mixing of signals.

  • Crosstalk: Interference from adjacent channels. Mitigation: Shielding, twisted pairs.

• Mitigation Techniques: Amplifiers (boost signal), equalizers (compensate distortion), shielding (block EMI).

H. Line Encoding Techniques

[!TIP] Manchester encoding embeds clock signal, eliminating need for separate clock line.

II. NETWORK ARCHITECTURE MODELS

A. OSI Reference Model

Layer Interaction & Encapsulation:

Data from upper layers is encapsulated with headers/trailers at each layer:

Application Data → Segment (L4) → Packet (L3) → Frame (L2) → Bits (L1).

B. TCP/IP Reference Model

Comparison with OSI:

• TCP/IP combines OSI's physical/data link into one layer.

• TCP/IP combines session/presentation/application into one layer.

• TCP/IP is protocol-centric (practical), OSI is model-centric (theoretical).

III. SWITCHING TECHNIQUES

A. Circuit Switching

• Dedicated physical path established between sender/receiver before data transfer.

• Phases: Setup → Data Transfer → Teardown.

• Example: PSTN telephone network.

• Advantages: Guaranteed bandwidth, low propagation delay.

• Disadvantages: Inefficient for bursty traffic, setup delay, resource reservation even during idle periods.

B. Packet Switching

• Datagram Approach (e.g., IP):

  • No pre-established path; each packet routed independently.

  • Packets may take different routes, arrive out-of-order.

  • Connectionless, stateless.

• Virtual Circuit Approach (e.g., Frame Relay, ATM):

  • Connection-oriented: path established before data transfer.

  • All packets follow same path, in-order delivery.

  • Resources reserved along path.

  • Types: PVC (permanent), SVC (switched).

C. Message Switching

• Store-and-forward entire message at each node.

• Advantages: Efficient use of bandwidth, can prioritize messages, supports heterogeneous networks.

• Disadvantages: High delay (store + forward), requires large storage, not suitable for real-time.

D. Cell Switching (ATM)

• Fixed 53-byte cells (5-byte header, 48-byte payload).

• Uses virtual circuits (VPI/VCI identifiers).

• Switching: Hardware-based fast cell switching via lookup tables.

• QoS Classes: CBR (constant bit rate), rt-VBR (real-time variable), nrt-VBR, ABR (available bit rate), UBR (unspecified).

• High speed, low delay, supports multimedia.

IV. DATA LINK LAYER

A. Framing

• Purpose: Define frame boundaries, allow error detection (FCS).

• Character-Oriented (Byte/Flag Stuffing):

  • Uses special flag bytes (e.g., 0x7E in PPP).

  • Byte stuffing: Insert escape byte (0x7D) when flag or escape byte appears in data.

• Bit-Oriented (Bit Stuffing):

  • Uses bit pattern (01111110) as delimiter (HDLC).

  • Bit stuffing: Insert 0 after five consecutive 1s in data to avoid delimiter pattern.

• Length-Based: Frame length field in header. Simple but error-prone if length corrupted.

[!TIP] Framing enables receiver to identify start/end of frames and extract data correctly.

B. Error Detection and Correction

Types of Errors:

• Single-bit error: One bit flipped (rare for high-speed).

• Burst error: Contiguous bits flipped (common due to noise).

Error Detection Methods:

1. Parity Check:

   • Vertical: Single parity bit per data unit.

   • Horizontal: Parity per row in block.

   • LRC (Longitudinal Redundancy Check): Block parity column-wise.

     Example: For data block:

     
     1011001
     
     1100101
     
     0110110
     
     

     Compute column-wise parity (XOR):

     Column 1: 1⊕1⊕0 = 0, Column 2: 0⊕1⊕1 = 0, ... → LRC = 0101101.

     Transmit: Original data + LRC row. Receiver recomputes; non-zero indicates error.

2. Checksum:

   • Sender: Divide data into 16-bit segments, sum them, take 1's complement.

   • Receiver: Sum all segments (including checksum); result should be 0.

   • Used in IP/TCP headers.

3. Cyclic Redundancy Check (CRC):

   • Polynomial division: Data as dividend, generator polynomial G(x) as divisor.

   • Append remainder (CRC bits) to data.

   • Receiver divides by same G(x); remainder 0 → no error.

   • Example: CRC-8 polynomial $$\displaystyle G(x) = x^8 + x^2 + x + 1 $$ (binary 100000111).

     Data: 11010101. Perform binary division → remainder (CRC).

     [!TIP] CRC detects all burst errors ≤ length of divisor and all odd-numbered errors if G(x) contains factor (x+1).

Error Correction Methods:

• Forward Error Correction (FEC): Add redundant bits; receiver corrects without retransmission (e.g., Hamming code).

• Automatic Repeat Request (ARQ):

  • Stop-and-Wait: Send one frame, wait for ACK. Utilization: $$\displaystyle \frac{1}{1+2a} $$ where $$\displaystyle a = \frac{\text{propagation time}}{\text{transmission time}} $$.

  • Sliding Window:

    • Go-Back-N: On error, retransmit from erroneous frame onward. Receiver discards subsequent frames.

    • Selective Repeat: Retransmit only erroneous frames; receiver buffers out-of-order frames.

C. Data Link Protocols

High-Level Data Link Control (HDLC):

• Frame Structure:

  
  Flag (01111110) | Address | Control | Information | FCS (CRC) | Flag
  
  

• Modes of Operation:

  • Normal Response Mode (NRM): Primary/secondary stations (used on multipoint).

  • Asynchronous Response Mode (ARM): Secondary can transmit without permission.

  • Asynchronous Balanced Mode (ABM): Both stations equal (most common, point-to-point).

• Features: Full-duplex, error control (ARQ), supports both connection-oriented/connectionless.

• Applications: Legacy WANs, PPP (derived from HDLC).

Other Protocols:

• X.25: Packet-switched, error correction at data link and network layers (slow, high overhead).

• Frame Relay: Simplified X.25, no error correction, only detection. Uses DLCI for virtual circuits.

V. LOCAL AREA NETWORKS (LANs)

A. Network Topologies

B. Ethernet (IEEE 802.3)

• CSMA/CD Mechanism:

  1. Carrier Sense: Listen before transmit.

  2. Collision Detection: Monitor while transmitting.

  3. Collision Handling: On collision, send jam signal, then backoff (binary exponential).

  4. Retransmission: After random backoff.

• Frame Structure:

  • Preamble (7 bytes 10101010 + 1 byte 10101011) for synchronization.

  • Destination MAC (6 bytes), Source MAC (6 bytes).

  • Type/Length (2 bytes): Indicates upper-layer protocol or frame length.

  • Data (46–1500 bytes), Pad (if needed).

  • FCS (4 bytes) – CRC-32.

C. Token-Based LANs

• Token Bus (IEEE 802.4):

  • Physical bus, logical ring.

  • Stations logically ordered by address; token passed in address order.

  • Used in industrial networks.

• Token Ring (IEEE 802.5):

  • Physical star (hub), logical ring.

  • Token (3-byte frame) circulates; station with token can transmit.

  • Uses priority bits and reservation field.

  • Speed: 4/16 Mbps.

D. Fiber Distributed Data Interface (FDDI)

• Dual-Ring Architecture: Primary ring (data) and secondary ring (backup).

• Token Passing: Station with token transmits; token released after frame return.

• Fault Tolerance: If primary ring breaks, stations "wrap" to secondary ring (ring wrap). Dual-homing for critical nodes.

• Speed: 100 Mbps over fiber.

• Applications: Backbone for high-speed LANs, campus networks.

VI. METROPOLITAN AND WIDE AREA NETWORKS

A. Distributed Queue Dual Bus (DQDB) – IEEE 802.6

• Architecture: Two unidirectional buses (A and B) in opposite directions.

• Operation:

  • Request for slot on bus A (upstream), data transmitted on bus B (downstream) or vice versa.

  • Each station maintains distributed queue for each bus.

  • Fair access: Requests queued in order of arrival; slots allocated fairly.

• MAN Support: High-speed (up to 155 Mbps), supports both connection-oriented and connectionless services.

B. Switched Multi-megabit Data Service (SMDS)

• Connectionless, datagram service for MAN/WAN.

• Uses 53-byte cells (similar to ATM) for switching.

• High-speed, supports bursty data.

• Largely superseded by Ethernet and MPLS.

C. Frame Relay

• Virtual Circuits: PVC (pre-configured), SVC (dynamic setup).

• Efficiency: No error correction (only detection via FCS), no flow control. Relies on higher layers for reliability.

• Frame Structure: Flag, Address (DLCI), Control, Information, FCS, Flag.

• Limitations: No congestion control (can lead to packet loss), best for reliable networks.

D. Asynchronous Transfer Mode (ATM)

• Cell-Based Architecture: Fixed 53-byte cells (5-byte header, 48-byte payload).

• Header Fields: VPI (Virtual Path Identifier), VCI (Virtual Circuit Identifier), PT (Payload Type), CLP (Cell Loss Priority), HEC (Header Error Control).

• Switching: Based on VPI/VCI lookup in switching fabric; fast hardware switching.

• QoS: Classes – CBR (voice), rt-VBR (video), nrt-VBR (data), ABR (congestion feedback), UBR (best-effort).

E. Synchronous Optical Network (SONET)

• Frame Structure: STS-1 frame (810 bytes, 125 µs). Transport overhead, payload.

• Hierarchy: STS-1 (51.84 Mbps), STS-3c (155.52 Mbps), STS-12c (622.08 Mbps), etc.

• Applications: Backbone networks, fiber optic links, carrier networks.

VII. NETWORKING DEVICES

[!TIP] Switches operate at Layer 2, routers at Layer 3. Switches reduce collision domains; routers separate broadcast domains.

VIII. ROUTING ALGORITHMS

A. Distance Vector Routing

• Bellman-Ford Algorithm: Each router maintains distance vector (cost to each destination). Periodically exchange entire vector with neighbors. Update:

  $$\displaystyle D_x(y) = \min_{v \in \text{neighbors}} \left[ c(x,v) + D_v(y) \right] $$

  where $$\displaystyle D_x(y) $$ = cost from $x$ to $y$, $c(x,v)$ = cost to neighbor $v$.

• Routing Table Updates: Every $n$ seconds (e.g., RIP every 30 sec).

• Count-to-Infinity Problem: When link fails, routers increment metric indefinitely.

  Solutions: Split horizon, route poisoning, hold-down timers.

• Example: RIP uses hop count (max 15 hops).

B. Link State Routing

• Dijkstra's Algorithm: Each router has complete map (link state database). Compute shortest path tree from itself to all destinations.

  1. Initialize: Current node = self, cost = 0; others = ∞.

  2. Select node with smallest tentative cost.

  3. Update costs to neighbors via selected node.

  4. Repeat until all nodes visited.

• Link State Advertisements (LSAs): Each router floods LSAs (describing its links) to all others; databases synchronized.

• Comparison with Distance Vector:

  | Aspect | Distance Vector | Link State | |---------------------|-----------------------------------|------------------------------------| | Information shared | Entire distance vector | Link state (neighbors only) | | Convergence speed | Slow (count-to-infinity) | Fast (flooding + Dijkstra) | | Memory requirement | Low (only neighbor vectors) | High (full map) | | Overhead | Periodic full updates | Event-triggered LSAs | | Example | RIP, IGRP | OSPF, IS-IS |

IX. STANDARDS AND ORGANIZATIONS

A. Importance of Standards

• Ensure interoperability between vendors.

• Promote competition, avoid monopolies.

• Provide common framework for development.

B. Key Standardization Bodies

X. ADDITIONAL TOPICS FROM PAST PAPERS

A. Overview of TCP/IP

• Protocol Suite: IP (network), TCP/UDP (transport), HTTP/FTP/SMTP (application).

• Key Features: Connectionless (IP), reliable (TCP), best-effort (UDP), hierarchical addressing.

• Importance: Foundation of Internet, widely deployed.

B. Network Classification

• Covered in Section I.C.

C. Error Control Mechanisms in Digital Broadcast (e.g., DBS-TV)

• FEC: Reed-Solomon codes, convolutional codes (Viterbi decoding).

• ARQ: Limited use due to latency; often hybrid ARQ (FEC + ARQ).

• Interleaving: Spread burst errors to appear as single errors for FEC.

• Importance: Maintains broadcast quality despite channel noise.

D. Comparison of DQDB and SMDS

E. Role and Functions of Internetworking Devices

• Covered in Section VII.

[!TIP] For exams, focus on OSI/TCP-IP models, HDLC, CRC/LRC calculations, switching comparisons, and routing algorithms (Bellman-Ford vs Dijkstra). Practice frame structures (Ethernet, HDLC) and error detection examples.