11.2 Data Communications, OSI Model & Protocols

11.2 Data Communications, OSI Model & Protocols

Quick Answer: Modern data networks structure communication tasks into functional layers. Key concepts heavily tested on the PRC ECE board exam include the 7 layers of the OSI model, TCP vs. UDP protocol characteristics, Cyclic Redundancy Check (CRC) polynomial division modulo-2, Automatic Repeat Request (ARQ) flow control, and IPv4 CIDR subnetting math ($2^{32-n} - 2$ usable host addresses per subnet).

---

Introduction to Data Communications & Network Architectures

Data communications refers to the exchange of digitized information between two or more nodes via a transmission medium governed by agreed-upon communication rules designated as protocols. To reduce system complexity, modern networking protocols are organized as a vertical stack of functional layers, where each layer provides specific services to the layer immediately above it while hiding internal implementation details.

---

The OSI 7-Layer Reference Model

Promulgated by the International Organization for Standardization (ISO) in 1984 under standard ISO/IEC 7498-1, the Open Systems Interconnection (OSI) model defines a 7-layer framework for network intercommunication.

  ┌─────────────────────────────────────────────────────────────┐
  │ Layer 7: Application  (HTTP, FTP, SMTP, DNS, SSH)           │
  ├─────────────────────────────────────────────────────────────┤
  │ Layer 6: Presentation (Data Formatting, Encryption, JPEG)  │
  ├─────────────────────────────────────────────────────────────┤
  │ Layer 5: Session      (Dialog Control, RPC, NetBIOS)        │
  ├─────────────────────────────────────────────────────────────┤
  │ Layer 4: Transport    (TCP, UDP, Segment/Datagram Flow)    │
  ├─────────────────────────────────────────────────────────────┤
  │ Layer 3: Network      (IPv4, IPv6, ICMP, OSPF - Routing)    │
  ├─────────────────────────────────────────────────────────────┤
  │ Layer 2: Data Link    (Ethernet, MAC/LLC, Switches, CRC)   │
  ├─────────────────────────────────────────────────────────────┤
  │ Layer 1: Physical    (RS-232, RJ-45, Cables, Bit Encoding) │
  └─────────────────────────────────────────────────────────────┘

Functional Overview of the 7 Layers

1. Layer 1 - Physical Layer: Transmits raw, unstructured bitstreams over physical mediums (copper wires, fiber optics, radio frequencies). Defines electrical voltage levels, bit timings, connector pinouts (e.g., RS-232, RJ-45), and physical data rates.
2. Layer 2 - Data Link Layer: Handles physical node-to-node framing, Media Access Control (MAC) addressing, error detection via CRC, and link-level flow control. Divided into two sublayers: Logical Link Control (LLC) (IEEE 802.2) and Media Access Control (MAC) (IEEE 802.3 Ethernet).
3. Layer 3 - Network Layer: Manages logical network addressing (IPv4/IPv6 addresses), packet routing across interconnected networks (routers), and packet fragmentation.
4. Layer 4 - Transport Layer: Manages end-to-end process-to-process communication, connection establishment, error recovery, segment sequencing, and port addressing. Key protocols: TCP and UDP.
5. Layer 5 - Session Layer: Establishes, maintains, synchronizes, and terminates session dialogs between endpoints (e.g., Remote Procedure Calls - RPC, NetBIOS).
6. Layer 6 - Presentation Layer: Translates data formats between application syntax and standard wire network syntax. Handles data compression, character encoding (ASCII, Unicode), and encryption/decryption (SSL/TLS).
7. Layer 7 - Application Layer: Provides direct network services to software applications (browsers, email clients). Examples: HTTP/HTTPS, FTP, SMTP, DNS, SNMP.

---

OSI Model vs. TCP/IP Architecture Comparison

While OSI serves as the formal theoretical reference model, the practical internet architecture is built upon the 4-layer (or 5-layer) TCP/IP Protocol Suite.

---

Transport Layer Protocols: TCP vs. UDP

The Transport Layer provides process-to-process data delivery using 16-bit Port Numbers (allowing $65,536$ distinct logical ports).

         TCP 3-Way Handshake                   UDP Connectionless
   Client                   Server       Client                   Server
     │                         │           │                         │
     │──── SYN (Seq=x) ───────>│           │──── Datagram ──────────>│
     │<─── SYN-ACK (Ack=x+1) ──│           │  (No Handshake/Ack)     │
     │──── ACK (Ack=y+1) ─────>│           │                         │

Transmission Control Protocol (TCP)

• Connection-Oriented: Establishes a reliable connection using a 3-Way Handshake (SYN $\rightarrow$ SYN-ACK $\rightarrow$ ACK).
• Reliability: Guarantees ordered data delivery using sequence numbers, positive acknowledgments (ACK), and automatic retransmission of lost segments.
• Flow Control & Congestion Control: Implements a dynamic Sliding Window Protocol to prevent receiver buffer overflow and congestion window mechanisms (Slow Start, Congestion Avoidance).
• Header Size: 20 bytes minimum (up to 60 bytes with options).

User Datagram Protocol (UDP)

• Connectionless: Transmits independent datagrams without establishing prior connections or performing handshakes.
• Unreliable & Low Overhead: Provides no sequence numbering, no acknowledgments, and no retransmission. Unordered datagrams may be lost or duplicated.
• Header Size: Minimal 8-byte fixed header (Source Port, Destination Port, Length, Checksum).
• Use Cases: Real-time applications prioritizing speed over reliability: VoIP, video streaming, DNS queries, online gaming.

---

Error Detection & Control: CRC Polynomial Division

Cyclic Redundancy Check (CRC) is a powerful error-detecting mechanism operating at Layer 2. A block of data bits is treated as coefficients of a binary polynomial and divided modulo-2 by a agreed-upon generator polynomial $G(X)$.

Mathematical Principles

1. Given an $k$-bit message polynomial $M(X)$ and an $(n+1)$-bit generator polynomial $G(X)$ of degree $n$.
2. Append $n$ zero bits to the message stream ($M(X) \cdot X^n$).
3. Divide $M(X) \cdot X^n$ by $G(X)$ using binary modulo-2 division (exclusive-OR / XOR operations without carry/borrow).
4. The remainder $R(X)$ of length $n$ bits is appended to the message as the CRC checksum field.
5. The receiver divides the received frame by $G(X)$. A remainder of zero indicates no errors occurred during transmission.

---

Step-by-Step CRC Calculation Example

Problem: Compute the 3-bit CRC remainder for a binary message $M = 110101$ using generator polynomial $G(X) = X^3 + X + 1$ (binary divisor 1011).

Solution:

1. Determine Generator Degree ($n$): $G(X) = 1X^3 + 0X^2 + 1X^1 + 1X^0 \rightarrow$ Divisor = 1011 ($n = 3$ bits).
2. Append $n=3$ Zeros to Message: Augmented Dividend = 110101000.
3. Perform Modulo-2 Division (XOR Subtraction):

                 111001  <-- Quotient (Discarded)
           ───────────────
    1011 ┘ 110101000
           1011
           ────
            1100
            1011
            ────
             1111
             1011
             ────
              1000
              0000
              ────
              1000
              0000
              ────
              1000
              1011
              ────
               011  <-- 3-Bit CRC Remainder R(X)

1. Final Transmitted Frame: Transmitted bits = Message + CRC = 110101011.

---

Automatic Repeat Request (ARQ) Mechanisms

When a receiver detects a corrupted frame via CRC failure, ARQ protocols manage retransmission:

1. Stop-and-Wait ARQ: Transmitter sends 1 frame and pauses, waiting for an ACK before sending the next frame. Simple but inefficient over long propagation delay paths.
2. Go-Back-N ARQ: Transmitter sends multiple frames up to a window size $W$. If frame $k$ is lost or corrupted, receiver rejects all subsequent frames ($k+1, k+2, \dots$), forcing the transmitter to retransmit frame $k$ and all subsequent frames in the window.
3. Selective Repeat ARQ: Transmitter retransmits only the specific missing or corrupted frame. Receiver maintains a buffer to store out-of-order frames. Most bandwidth-efficient method.

---

IPv4 Addressing & CIDR Subnetting Guide

An IPv4 address is a 32-bit binary number represented in dotted-decimal notation (four 8-bit octets separated by dots, e.g., 192.168.1.1).

Classless Inter-Domain Routing (CIDR)

CIDR notation appends a slash followed by the prefix length $n$ indicating how many leading bits constitute the Network ID (e.g., 192.168.1.0/24).

$$\text{Total IP Addresses} = 2^{32-n}$$ $$\text{Usable Host IP Addresses} = 2^{32-n} - 2$$

(Two addresses are reserved: the Network Address with all host bits set to 0, and the Directed Broadcast Address with all host bits set to 1).

Subnetting Reference Table

Step-by-Step Subnetting Calculation Example

Problem: Given the IP address 192.168.10.130/27, determine (a) the Subnet Mask, (b) the Network ID, (c) the Broadcast Address, and (d) the range of usable host IP addresses.

Solution:

1. Subnet Mask: /27 prefix means 27 ones followed by 5 zeros $\rightarrow$ 11111111.11111111.11111111.11100000 = 255.255.255.224.
2. Block Size (Magic Number): $256 - 224 = 32$ (or $2^5 = 32$). Subnet multiples: 0, 32, 64, 96, 128, 160...
3. Locate Subnet for Host IP .130: The host IP .130 lies between 128 and 159.
   • Network ID: 192.168.10.128
   • Broadcast Address: 192.168.10.159 (one less than next subnet 160)
   • Usable Host Range: 192.168.10.129 to 192.168.10.158 (30 usable hosts).