CS8591

Computer Networks

7 Files
•
Updated Jan 2026
Download Progress 0/7 files
Unit 1
Notes Not downloaded 1.4 MB
Unit 2
Notes Not downloaded 1.1 MB
Unit 3
Notes Not downloaded 1.2 MB
Unit 4
Notes Not downloaded 966 KB
Unit 5
Notes Not downloaded 715 KB
Important Questions
Study Not downloaded 372 KB
Question Paper
Exam Not downloaded 558 KB
Academic Study Guide

Computer Networks Complete Academic Curriculum & Revision Guide

Computer Networks enable distributed communication across autonomous systems. This academic guide provides in-depth conceptual notes on the 7-layer OSI and TCP/IP protocol suites, data link layer framing and error control, IPv4/IPv6 addressing and subnetting, routing protocols (RIP, OSPF, BGP), transport-layer flow and congestion control (TCP), and core application protocols (DNS, HTTP, SMTP).

Unit-by-Unit Theoretical Revision

Unit 1: Network Architectures & Physical Layer Fundamentals

Computer networks organize communication functions into layered architectures where each layer provides well-defined services to the layer above it.

  • OSI 7-Layer Model: Physical (bits) → Data Link (frames) → Network (packets) → Transport (segments) → Session (dialogues) → Presentation (syntax/encryption) → Application (user services).
  • TCP/IP Protocol Suite: 4-layer pragmatic architecture: Link Layer, Internet Layer (IP), Transport Layer (TCP/UDP), and Application Layer.
  • Channel Capacity Theorems:
    • Nyquist Theorem (Noiseless): Max Bit Rate = 2B · logâ‚‚(M) bps, where B is bandwidth in Hertz and M is signal discrete voltage levels.
    • Shannon Capacity Theorem (Noisy): Capacity = B · logâ‚‚(1 + SNR) bps, where SNR is signal-to-noise ratio in linear power.
  • Transmission Media: Guided media (Twisted Pair, Coaxial, Fiber Optic) vs Unguided media (Radio, Microwave, Infrared).

Unit 2: Data Link Layer & Medium Access Control (MAC)

  • Framing & Error Handling: Character count, byte stuffing (flag bytes), and bit stuffing (inserting 0 after five consecutive 1s).
  • Error Detection & Correction:
    • Cyclic Redundancy Check (CRC): Uses polynomial division using XOR arithmetic to generate Frame Check Sequence (FCS) remainder bits.
    • Hamming Code: Forward Error Correction (FEC) computing parity bits at positions 2^k to detect and correct single-bit errors.
  • Sliding Window Flow Control Protocols:
    • Stop-and-Wait ARQ: Sender transmits one frame and waits for ACK. Efficiency η = 1 / (1 + 2a), where a = T_prop / T_trans.
    • Go-Back-N ARQ: Sender window size W = 2^k - 1, receiver window size = 1. Retransmits all frames starting from damaged frame.
    • Selective Repeat ARQ: Sender and receiver windows equal 2^(k-1). Receives out-of-order frames and selectively retransmits only corrupted frames.
  • MAC Protocols: Pure ALOHA (max efficiency 18.4%), Slotted ALOHA (efficiency 36.8%), CSMA/CD (Ethernet collision detection with binary exponential backoff), and CSMA/CA (Wi-Fi collision avoidance with RTS/CTS).

Unit 3: Network Layer & Routing Protocols

  • IPv4 Addressing & Subnetting: 32-bit addresses. Classful addressing (Classes A, B, C, D, E) superseded by Classless Inter-Domain Routing (CIDR) using subnet masks (e.g., 192.168.1.0/24).
  • IPv6 Architecture: 128-bit hexadecimal addressing format eliminating NAT requirements, with simplified fixed 40-byte base header and auto-configuration.
  • Routing Algorithms:
    • Distance Vector Routing (Bellman-Ford): Nodes exchange routing vectors with immediate neighbors. Suffers from Count-to-Infinity problem; resolved by Split Horizon and Poison Reverse. (Used in RIP).
    • Link State Routing (Dijkstra’s Algorithm): Every node floods link state packets (LSPs) across the entire network and computes shortest paths independently. (Used in OSPF).
    • Path Vector Routing: Exchanges complete AS path attributes to enforce policy-based inter-domain routing. (Used in BGP).
  • Auxiliary Network Protocols: ARP (resolves IP to MAC address), ICMP (network diagnostics, ping, traceroute), and DHCP (dynamic IP allocation).

Unit 4: Transport Layer & TCP Congestion Control

  • TCP vs UDP: TCP provides connection-oriented, reliable, byte-stream delivery with flow and congestion control. UDP provides connectionless, lightweight, unreliable datagram delivery for low-latency streaming and DNS.
  • TCP 3-Way Handshake: Connection establishment: Client sends SYN → Server responds with SYN + ACK → Client sends ACK. Teardown uses 4-way FIN handshakes.
  • TCP Congestion Control Mechanisms:
    • Slow Start: Congestion Window (cwnd) doubles exponentially every RTT until reaching Slow Start Threshold (ssthresh).
    • Congestion Avoidance: cwnd grows linearly by 1 MSS per RTT upon reaching ssthresh (Additive Increase).
    • Fast Retransmit & Fast Recovery (TCP Reno): 3 duplicate ACKs trigger immediate retransmission of missing segment; ssthresh set to cwnd/2 and cwnd set to ssthresh + 3 MSS without dropping to 1 MSS.

Unit 5: Application Layer & Network Security

  • Domain Name System (DNS): Hierarchical distributed database mapping hostnames to IP addresses using root servers, TLD servers, and authoritative name servers over UDP port 53.
  • Hypertext Transfer Protocol (HTTP): HTTP/1.1 (persistent TCP connections, pipelining), HTTP/2 (binary framing, multiplexing over single TCP connection, server push), and HTTP/3 (QUIC protocol over UDP).
  • Electronic Mail: SMTP (push protocol, port 25), POP3 (pull protocol, deletes from server), and IMAP (pull protocol, syncs mailbox across devices).
  • Cryptography & Security: Symmetric encryption (AES, DES) vs Asymmetric encryption (RSA public-key cryptography), Digital Signatures (SHA-256 with private key), and TLS/SSL handshake.

Core Protocols Reference Matrix

Protocol Layer Underlying Transport Default Port Primary Function
DNSApplicationUDP / TCP53Resolves hostnames to IP addresses
HTTP / HTTPSApplicationTCP80 / 443Hypertext & REST web content transfer
DHCPApplicationUDP67 (Server) / 68 (Client)Automated IP configuration leasing
OSPFNetwork / RoutingIP (Protocol 89)N/AIntra-domain link-state routing
BGPApplication / RoutingTCP179Inter-domain path-vector routing
ARPData Link / NetworkEthernet BroadcastN/ATranslates IP to physical MAC address

High-Frequency University Exam Questions

Q1: Differentiate between Go-Back-N and Selective Repeat ARQ protocols.

In Go-Back-N ARQ, the receiver window size is strictly 1; if a frame is corrupted, the receiver silently discards it and all subsequent out-of-order frames, forcing the sender to retransmit the entire window of N frames. In Selective Repeat ARQ, the receiver window size is equal to sender window size (2^(n-1)); it buffers valid out-of-order frames and sends negative acknowledgments (NAKs) for corrupted frames, prompting the sender to retransmit only the damaged frame.

Q2: Explain the Count-to-Infinity problem in Distance Vector Routing and how Split Horizon resolves it.

The Count-to-Infinity problem occurs when a network link fails, but neighboring routers continue updating each other with stale, incrementing metric values due to slow convergence in Bellman-Ford calculations. Split Horizon solves this by enforcing a rule that routing information about a destination is never advertised back out of the interface through which it was learned.

Q3: Describe the TCP 3-Way Handshake mechanism for establishing a reliable connection.

TCP establishes connections using three messages: (1) SYN: The client selects an initial sequence number x and transmits a SYN segment. (2) SYN-ACK: The server acknowledges receipt with ACK = x + 1, chooses its own initial sequence number y, and sends a SYN-ACK. (3) ACK: The client responds with an ACK segment containing seq = x + 1 and ack = y + 1. Data transmission can commence immediately following this exchange.