Complete Computer Networks Notes for BSc CSIT 4th Semester (TU) | OSI, TCP/IP, Subnetting, Routing, Past Questions & Exam Guide

Introduction

Computer Networks (CSC263) shows up in the 4th semester of BSc CSIT and immediately becomes one of those subjects where students either feel genuinely interested or genuinely lost — sometimes both at the same time. It has a reputation for being conceptually heavy: you're dealing with seven-layer models, IP addressing math, routing algorithms with graphs, protocol comparisons, and terms that blur together (TCP vs UDP, OSI vs TCP/IP, Distance Vector vs Link State) until you've sat with them long enough.

What makes CN worth taking seriously beyond just passing the exam is that almost everything happening when your device connects to anything — loading a webpage, sending a message, streaming video — is the protocols in this course doing their jobs. When a DNS query resolves a hostname, that's Unit 6. When TCP retransmits a lost packet, that's Unit 5. When your packet hops across routers to reach its destination, that's Unit 4. You're not learning abstract theory; you're learning what the internet is actually made of.

For exam purposes, CN is one of those subjects where the same topics come up year after year with minor variation: OSI model, TCP header, subnetting calculations, Dijkstra's algorithm, congestion control, DNS resolution. This guide teaches each one properly rather than just listing them, so you can answer variations you haven't seen before rather than only the exact questions you've memorized.

Official Course Information

Particular	Details
Course Title	Computer Networks
Course Code	CSC263
Semester	Fourth Semester
Nature of Course	Theory + Lab
Full Marks	60 + 20 + 20
Pass Marks	24 + 8 + 8
Credit Hours	3

Course Description and Objective

CSC263 introduces the concept of computer networking and examines the different layers of the networking model. The course objective is to develop a working understanding of computer networking — its layers, topologies, protocols and standards, IPv4/IPv6 addressing, routing, and the latest networking standards. The lab component is genuinely practical: it includes packet capture and analysis using Wireshark, building and testing networks in Packet Tracer, configuring routing protocols (RIP, OSPF, BGP), firewall and ACL setup, and a case-study visit to a real Network Operations Center.

Official Unit-Wise Syllabus Overview

Unit	Title	Hours
1	Introduction to Computer Networks	6
2	Physical Layer and Network Media	4
3	Data Link Layer	8
4	Network Layer	10
5	Transport Layer	6
6	Application Layer	7
7	Multimedia and Future Networking	4

Unit 4 (Network Layer) carries the most teaching hours and is consistently the heaviest unit in exams — subnetting and routing together almost always account for multiple marks. Unit 3 (Data Link Layer) and Unit 6 (Application Layer) are both 8 and 7 hours respectively and appear regularly. Unit 7 covers SDN and future networking, which is a common short-question topic even though the unit is relatively brief.

---

Unit 1: Introduction to Computer Networks

A computer network is a collection of devices interconnected by communication links that can share resources and exchange data. Understanding why networks exist matters more than memorizing the definition: without networking, every file would have to be physically transferred, every service would be isolated, and the idea of "the internet" simply wouldn't exist.

Network Topologies

Topology	Description	Advantage	Disadvantage
Star	All nodes connect to a central hub/switch	Easy to add/remove devices, failure of one node doesn't affect others	Central device failure brings down entire network
Bus	All nodes share a single backbone cable	Simple and inexpensive for small networks	Entire network down if backbone fails, collisions common
Ring	Nodes connected in a closed loop	Predictable performance, no collisions in token ring	Single point of failure; adding devices disrupts the ring
Tree	Hierarchical star topology	Scalable, easy to manage in segments	Root node failure affects entire network
Mesh	Every node connected to every other	Highly fault-tolerant, multiple paths	Expensive, complex cabling

Network Types

Type	Coverage	Example
PAN	Personal area (~10m)	Bluetooth phone-to-headset
LAN	Building or campus	Office network, college lab
CAN	Campus-wide	University network across buildings
MAN	City-wide	City ISP network
WAN	Country or global	Internet

The OSI Reference Model

The OSI (Open Systems Interconnection) model is a seven-layer framework developed by ISO to standardize how different systems communicate. No actual protocol perfectly implements the OSI model — it's a conceptual reference — but it's the language used to describe where a given protocol or device operates.

Layer	Name	PDU	Devices	Key Protocols	Function
7	Application	Data	—	HTTP, FTP, DNS, SMTP, DHCP	Interface between user/software and network
6	Presentation	Data	—	SSL/TLS, JPEG, ASCII	Data formatting, encryption, compression
5	Session	Data	—	NetBIOS, RPC	Establishes, manages, terminates sessions
4	Transport	Segment	—	TCP, UDP	End-to-end delivery, error recovery, flow control
3	Network	Packet	Router	IP, ICMP, RIP, OSPF	Logical addressing, routing between networks
2	Data Link	Frame	Switch, Bridge	Ethernet, PPP, HDLC	Physical addressing (MAC), error detection
1	Physical	Bits	Hub, Repeater	Ethernet cable, Wi-Fi	Transmits raw bits over a physical medium

Memory trick for layers (top to bottom): All People Seem To Need Data Processing (Application, Presentation, Session, Transport, Network, Data Link, Physical).

OSI vs TCP/IP Model

Aspect	OSI Model	TCP/IP Model
Layers	7	4
Layer structure	Application, Presentation, Session, Transport, Network, Data Link, Physical	Application, Transport, Internet, Network Access
Development	Developed before protocols were designed (theoretical first)	Developed around existing protocols (practical first)
Usage	Reference model for teaching and comparison	Actual model used by the internet
Session/Presentation	Separate layers	Merged into Application layer

---

Unit 2: Physical Layer and Network Media

The physical layer deals with the actual transmission of bits over a medium — electrical signals, light pulses, or radio waves. Nothing at this layer knows about packets, addresses, or protocols; it just moves raw bits.

Network Devices at Each Layer

Device	Layer	Function
Repeater	Physical	Amplifies/regenerates signal to extend range
Hub	Physical	Broadcasts signal to all connected ports
Bridge	Data Link	Forwards frames between network segments using MAC addresses
Switch	Data Link	Forwards frames only to the intended port (learned MAC table)
Router	Network	Forwards packets between different networks using IP addresses

Transmission Media

Wired media:

• Twisted Pair (UTP/STP): Pairs of copper wires twisted to reduce interference. Cat5e and Cat6 are common UTP standards supporting 100Mbps–10Gbps at limited distances.
• Coaxial Cable: Central copper conductor with shielding, used in older Ethernet and cable TV.
• Fiber Optic: Transmits light pulses through glass/plastic fiber. Immune to electromagnetic interference, much higher bandwidth and longer range than copper, but more expensive.

Wireless media:

• Radio Waves: Used in Wi-Fi, Bluetooth; can pass through walls, omnidirectional.
• Microwaves: Directional, used for long-distance point-to-point links and satellite communication.
• Infrared: Short range, line-of-sight only (remote controls, old IrDA connections).

Circuit, Message, and Packet Switching

Switching Type	How It Works	Example
Circuit Switching	A dedicated path is established for the entire duration of communication	Traditional telephone calls
Message Switching	Entire message stored at each node before being forwarded (store-and-forward)	Old telegraph systems
Packet Switching	Data broken into packets; each takes its own route and is reassembled at destination	The internet

Packet switching is why the internet is efficient: if one path gets congested, packets simply take another route, and multiple users can share the same links without reserving them entirely.

---

Unit 3: Data Link Layer

The Data Link Layer (DLL) bridges the gap between the physical transmission of bits and the logical network above it. Its two main sublayers are LLC (Logical Link Control), which interfaces with the network layer, and MAC (Media Access Control), which handles access to the physical medium and physical addressing.

Error Detection Techniques

Technique	How It Works
Parity Check	Adds one bit to make total 1s even (even parity) or odd (odd parity). Detects single-bit errors.
Checksum	Divides data into chunks, adds them; receiver checks that sum equals expected value.
CRC (Cyclic Redundancy Check)	Divides bit string by a generator polynomial; remainder sent with data. Very reliable.

Flow Control and ARQ Protocols

Flow control prevents a fast sender from overwhelming a slow receiver.

Stop-and-Wait ARQ: Sender sends one frame, then waits for an acknowledgment before sending the next. Simple but inefficient — the link is idle while waiting.

Go-Back-N ARQ: Sender can have up to N frames outstanding (unacknowledged). If one is lost or corrupted, the receiver discards that frame and all subsequent ones, and sender retransmits from the error frame. Window size = N.

Selective Repeat ARQ: Like Go-Back-N, but the receiver buffers correctly received frames even if an earlier one is missing. Only the specific lost frame gets retransmitted. More efficient but requires more receiver buffer space.

Channel Access Methods

ALOHA: Stations transmit whenever they have data. Collisions occur; wait a random time and retransmit.

Slotted ALOHA: Time is divided into slots; stations can only transmit at the start of a slot. Reduces collision probability compared to pure ALOHA.

CSMA/CD (802.3): "Listen before talking" — station checks if medium is idle before transmitting. If a collision is detected during transmission, all stations stop, wait a random backoff time, and retry. Used in traditional wired Ethernet.

CSMA/CA (802.11): Used in Wi-Fi. Collisions can't be reliably detected wirelessly, so CSMA/CA tries to avoid them by using backoff timers before transmission.

---

Unit 4: Network Layer

The Network Layer is the core of the entire course and the most heavily examined unit. It handles logical addressing, moving packets across multiple networks (routing), and managing the IP protocol that the internet runs on.

IPv4 Addressing

An IPv4 address is a 32-bit number written as four octets in dotted decimal notation (e.g., 192.168.1.10). The address identifies both the network a host belongs to and the host itself within that network.

Address Classes (Classful Addressing):

Class	Range	Default Subnet Mask	Usage
A	1.0.0.0 – 126.255.255.255	255.0.0.0 (/8)	Very large organizations
B	128.0.0.0 – 191.255.255.255	255.255.0.0 (/16)	Large organizations
C	192.0.0.0 – 223.255.255.255	255.255.255.0 (/24)	Small organizations
D	224.0.0.0 – 239.255.255.255	—	Multicast
E	240.0.0.0 – 255.255.255.255	—	Reserved/experimental

Subnetting — Solved Example

Subnetting divides a single network into smaller subnetworks to use address space more efficiently and reduce broadcast traffic.

Question: Given the network 192.168.1.0/26, find the Network ID, Broadcast Address, First Usable Host, Last Usable Host, and total number of hosts.

Solution:

Step 1: /26 means 26 bits are network bits, so 32 - 26 = 6 bits remain for hosts.

Step 2: Total addresses = 2⁶ = 64. Usable hosts = 64 - 2 = 62 (subtract network and broadcast).

Step 3: Subnet mask = 11111111.11111111.11111111.11000000 = 255.255.255.192

Step 4:

• Network ID: 192.168.1.0
• First Usable Host: 192.168.1.1
• Last Usable Host: 192.168.1.62
• Broadcast Address: 192.168.1.63

Second example — /27:

Network: 192.168.1.0/27 → 32 - 27 = 5 host bits → 32 addresses, 30 usable hosts

• Network ID: 192.168.1.0
• Broadcast: 192.168.1.31
• First Host: 192.168.1.1 | Last Host: 192.168.1.30

IPv4 vs IPv6

Aspect	IPv4	IPv6
Address length	32 bits	128 bits
Notation	Dotted decimal (192.168.1.1)	Hexadecimal with colons (2001:db8::1)
Total addresses	~4.3 billion	~3.4 × 10³⁸
Header size	20 bytes (variable)	40 bytes (fixed)
Fragmentation	Router or host can fragment	Only source host fragments
Security	Optional (IPSec)	Built-in IPSec support
NAT required	Yes (address exhaustion workaround)	Not needed
Configuration	Manual or DHCP	Auto-configuration (SLAAC)

Routing

Routing is the process of selecting paths in a network for traffic to travel from source to destination.

Type	Description
Static Routing	Routes configured manually by administrator; doesn't adapt to network changes
Dynamic Routing	Routers exchange information and automatically adapt to topology changes
Distance Vector	Each router knows distance to every other router; shares its table with neighbors (e.g., RIP)
Link State	Each router knows the full topology; independently computes shortest paths (e.g., OSPF)
Interior Gateway Protocol	Routing within a single autonomous system (RIP, OSPF)
Exterior Gateway Protocol	Routing between autonomous systems (BGP)

Dijkstra's Algorithm — Solved Example

Dijkstra's algorithm finds the shortest path from a source node to all other nodes.

Graph: Nodes A, B, C, D, E with edges: A-B: 4, A-C: 2, B-C: 1, B-D: 5, C-D: 8, C-E: 10, D-E: 2

Source: A

Step	Visited	A	B	C	D	E
Initial	{A}	0	∞	∞	∞	∞
Visit A	{A}	0	4	2	∞	∞
Visit C (min=2)	{A,C}	0	3	2	10	12
Visit B (min=3)	{A,C,B}	0	3	2	8	12
Visit D (min=8)	{A,C,B,D}	0	3	2	8	10
Visit E (min=10)	{A,C,B,D,E}	0	3	2	8	10

Shortest paths from A: A→B: 3 (via C), A→C: 2, A→D: 8 (via C→B→D), A→E: 10 (via D→E)

Bellman-Ford Algorithm: Unlike Dijkstra's, Bellman-Ford works with negative edge weights and can detect negative cycles. It relaxes all edges (V-1) times, where V is the number of vertices. If a shorter path is still found after V-1 iterations, a negative cycle exists. The algorithm is used as the basis for the RIP routing protocol.

---

Unit 5: Transport Layer

The Transport Layer provides end-to-end communication services for applications — it's the bridge between the application running on your device and the network infrastructure below.

TCP vs UDP

Aspect	TCP	UDP
Connection	Connection-oriented (handshake required)	Connectionless (no handshake)
Reliability	Guaranteed delivery, acknowledgments, retransmission	No guarantee; best-effort
Ordering	Segments delivered in order	No ordering guarantee
Flow Control	Yes (sliding window)	No
Congestion Control	Yes	No
Speed	Slower due to overhead	Faster
Header Size	20 bytes minimum	8 bytes
Use Cases	HTTP, FTP, SMTP, SSH	DNS queries, VoIP, video streaming, online gaming

TCP 3-Way Handshake (Connection Establishment)

The 3-way handshake establishes a TCP connection before any data is sent:

1. SYN: Client sends a segment with SYN flag set, includes its initial sequence number (ISN). "I want to connect; my starting sequence number is X."
2. SYN-ACK: Server responds with SYN and ACK flags set, acknowledges client's SYN (ACK = X+1), and sends its own ISN. "I accept; my starting sequence number is Y; I acknowledge your X."
3. ACK: Client acknowledges server's SYN (ACK = Y+1). "Acknowledged; connection established."

TCP 4-Way Termination (Connection Release)

Closing a TCP connection takes four steps because each direction closes independently:

1. Client sends FIN (finished sending)
2. Server sends ACK of client's FIN
3. Server sends its own FIN
4. Client sends ACK of server's FIN, then waits TIME_WAIT period before fully closing

Congestion Control

TCP congestion control prevents a sender from overwhelming the network. The key mechanisms:

Slow Start: TCP begins with a congestion window (cwnd) of 1 MSS and doubles it each round-trip time (exponential growth) until a threshold is reached.

Congestion Avoidance (AIMD): After the threshold, cwnd increases by 1 MSS per RTT (additive increase). On a loss event detected by timeout, cwnd is cut back dramatically (multiplicative decrease).

Fast Retransmit: If the sender receives three duplicate ACKs for the same segment, it retransmits that segment immediately without waiting for a timeout.

Fast Recovery: After fast retransmit, TCP reduces cwnd by half but doesn't drop all the way back to 1 — it stays in congestion avoidance rather than restarting slow start.

Traffic Shaping: Leaky Bucket vs Token Bucket

Both algorithms regulate the flow of packets into a network to smooth out bursts.

Aspect	Leaky Bucket	Token Bucket
Mechanism	Packets enter a bucket and exit at constant rate; excess is discarded or queued	Tokens accumulate at a fixed rate; each packet requires a token to be sent
Output Rate	Always constant (like water leaking at fixed rate)	Bursty — can use saved-up tokens to send burst up to bucket capacity
Burst Handling	Does not allow bursts	Allows controlled bursts
Use Case	Ensuring constant transmission rate	Allowing bursts within overall rate limits

---

Unit 6: Application Layer

The Application Layer is the one users actually interact with — it's where web browsing, email, file transfer, and name resolution live.

DNS — How Domain Name Resolution Works

DNS (Domain Name System) translates human-readable hostnames (like google.com) into IP addresses that routers can use to forward packets. The resolution process:

1. User types google.com in the browser.
2. Browser checks its local DNS cache — if found, done.
3. If not, query goes to the Recursive Resolver (usually your ISP's DNS server).
4. Resolver checks its cache — if found, returns result.
5. Resolver queries a Root Name Server — returns address of the .com TLD server.
6. Resolver queries the .com TLD Server — returns address of Google's authoritative name server.
7. Resolver queries Google's Authoritative Name Server — returns the IP address for google.com.
8. Resolver caches the result and returns the IP to the browser.
9. Browser connects to the IP address.

DNS Query Types:

• Recursive Query: Client asks resolver to do all the work and return a final answer.
• Iterative Query: Resolver asks each server in turn, getting a referral each time, until it reaches the authoritative server.

HTTP

HTTP (HyperText Transfer Protocol) is a stateless, application-layer protocol used for transferring web pages. Each request-response cycle is independent — the server doesn't remember anything about previous requests without additional mechanisms like cookies or sessions. HTTP/1.1 introduced persistent connections; HTTP/2 added multiplexing (multiple requests over a single connection).

Email Protocols

Protocol	Purpose	Port
SMTP	Sending email (client→server, server→server)	25 / 587
IMAP	Retrieves email, leaves a copy on server	143 / 993 (SSL)
POP3	Downloads email and removes from server	110 / 995 (SSL)

ARP (Address Resolution Protocol)

ARP resolves IP addresses to MAC addresses within a local network. When a host wants to send a packet to an IP address on the same LAN, it broadcasts an ARP Request ("Who has IP 192.168.1.5?"). The host with that IP responds with an ARP Reply ("I have it; my MAC is AA:BB:CC:DD:EE:FF"). The requesting host caches this mapping in its ARP table.

---

Unit 7: Multimedia and Future Networking

SDN — Software Defined Networking

Traditional networking tightly couples the control plane (routing decisions) and the data plane (actual packet forwarding) inside each individual router or switch. This makes the network hard to program, manage, and adapt to new requirements.

SDN separates these two planes:

Aspect	Traditional Networking	SDN
Control Plane	Distributed across every device	Centralized in an SDN Controller
Data Plane	Tightly coupled with control	Programmable forwarding elements (switches)
Configuration	Device-by-device (CLI/SNMP)	Centrally programmed via APIs
Flexibility	Low — changes require touching every device	High — software-defined rules pushed from controller
Example	Traditional routers, switches	OpenFlow-enabled switches, controllers like ONOS/ODL

The SDN controller has a global view of the network and can dynamically reconfigure forwarding rules across all switches, making it much easier to implement load balancing, traffic engineering, and security policies compared to the distributed, box-by-box approach of traditional networks.

NFV (Network Function Virtualization) takes this further: network functions like firewalls, load balancers, and NAT gateways that used to require dedicated hardware can be run as software on commodity servers.

---

Important Past-Year Long Questions (10/20 Marks)

These questions have appeared repeatedly in TU CN exams. Preparing each one properly with diagrams and examples gives you a strong chance of covering what actually appears.

1. Explain the OSI reference model with a neat diagram. Compare OSI with the TCP/IP model.
2. Explain the TCP protocol. Describe TCP header fields and the 3-way handshake process.
3. Explain IPv4 addressing. Differentiate between classful and classless addressing, with a subnetting example.
4. Compare Distance Vector and Link State routing. Explain Dijkstra's algorithm with a step-by-step solved example.
5. Explain congestion control techniques in TCP — Slow Start, AIMD, Fast Retransmit, and Fast Recovery.
6. Explain DNS architecture. Describe the iterative and recursive DNS query resolution process.
7. What is flow control? Explain Stop-and-Wait ARQ and Go-Back-N with diagrams.
8. Explain IP fragmentation and reassembly.
9. Explain different network topologies with advantages and disadvantages.
10. Explain SDN architecture. Compare with traditional networking.

Important Short Questions (5 Marks)

1. Define protocol. Why are network standards required?
2. Differentiate between guided and unguided transmission media.
3. Unicast vs Multicast vs Broadcast communication.
4. Difference between Port and Socket.
5. Leaky Bucket vs Token Bucket algorithm.
6. Explain ARP Request and ARP Reply.
7. IPv4 vs IPv6 differences.
8. Functions of the Presentation Layer.
9. What is a Backbone Network?
10. Infrared communication — characteristics and limitations.
11. Difference between TCP and UDP.
12. What is ICMP? What is it used for?
13. CSMA/CD vs CSMA/CA.
14. What is NAT? Why is it used?
15. Difference between SMTP, IMAP, and POP3.

---

Formula and Conversion Reference

Topic	Formula / Reference
Number of hosts in subnet	2ⁿ - 2 (where n = host bits)
Number of subnets	2ˢ (where s = subnet bits borrowed)
Subnet mask from prefix	Convert prefix length bits to binary, fill 1s for network bits
Bandwidth-Delay Product	Bandwidth × Round-Trip Time
Throughput (Stop-and-Wait)	(Frame size) / (Frame size + 2 × Propagation delay × Bandwidth)
Dijkstra time complexity	O(V²) basic; O(E log V) with priority queue
Bellman-Ford time complexity	O(V × E)

---

Exam Writing Strategy

Scoring high in Computer Networks at TU depends heavily on how you present answers, not just whether you know the content.

Diagrams are mandatory — OSI model, TCP header, subnetting table, routing graph. An answer about the OSI model without a layered diagram will lose marks even if the text is correct. For subnetting and routing algorithm questions, show every step clearly since TU marks each step, not just the final answer.

Structure every long answer: definition → diagram → detailed explanation → example. The examiner is reading many papers quickly; a well-structured answer with labeled diagrams reads faster and makes it easier to award marks.

Time management: approximately 3 minutes per mark. A 10-mark question deserves around 30 minutes. Don't sacrifice later questions for a perfect earlier answer.

---

Study Plans

30-Day Plan

• Days 1–5: Unit 1 — OSI model (memorize every layer's details), TCP/IP comparison, topologies, network types
• Days 6–9: Unit 2 — physical layer, transmission media, switching types
• Days 10–14: Unit 3 — data link layer, ARQ protocols, error detection, ALOHA, CSMA/CD
• Days 15–21: Unit 4 — IPv4 addressing, subnetting (practice daily), IPv6, Dijkstra's, Bellman-Ford, routing protocols (RIP, OSPF, BGP)
• Days 22–25: Unit 5 — TCP header, 3-way handshake, congestion control, Leaky/Token Bucket
• Days 26–28: Unit 6 — DNS resolution, HTTP, email protocols (SMTP/IMAP/POP3), ARP, SNMP
• Days 29–30: Unit 7 (SDN, NFV) + full past-question review + OSI and subnetting final drill

15-Day Plan

• Days 1–2: Unit 1 (OSI model + TCP/IP comparison)
• Days 3–4: Units 2–3
• Days 5–8: Unit 4 (subnetting and routing deserve 4 full days)
• Days 9–11: Unit 5 (TCP, congestion control)
• Days 12–13: Unit 6 (DNS, HTTP, email, ARP)
• Day 14: Unit 7 + past long questions review
• Day 15: All comparison tables + one subnetting + one routing algorithm practice

7-Day Plan

• Day 1: Unit 1 (OSI table — learn it completely)
• Day 2: Units 2–3
• Day 3–4: Unit 4 (subnetting and Dijkstra — don't compress)
• Day 5: Unit 5
• Day 6: Units 6–7
• Day 7: Past questions + comparison tables + subnetting drill

Night-Before Exam

• Redo one subnetting problem and one Dijkstra table from scratch without looking — these are always on the exam and both require a clear head, not cramming.
• Skim the OSI layer table once — if you can't recall a layer's device/protocol/PDU quickly, that's a short question you're giving away.
• Review TCP vs UDP and Leaky Bucket vs Token Bucket comparison tables.
• Don't attempt new material — review only what you've already practiced.

---

Common Mistakes Students Make

Subnetting arithmetic errors are the most common marks-loser — students subtract the wrong number or confuse network ID and broadcast address. Work every subnetting problem as a table (total addresses → network ID → broadcast → first host → last host) rather than doing it in your head.

The OSI model is memorized rather than understood: students can list the layers but mix up which devices operate at which layer (a hub is Physical, a switch is Data Link, a router is Network — confusing these is an immediate mark deduction). The TCP 3-way handshake gets confused with the 4-way termination — remember that connection establishment is 3 steps and termination is 4, because each direction needs its own FIN and ACK. Dijkstra's algorithm gets set up wrong: students forget to initialize all distances to infinity, or pick the wrong minimum at each step. And TCP vs UDP gets regurgitated as a list without understanding why UDP is used for streaming despite being unreliable — the answer is that retransmission delays are worse for real-time media than occasional dropped frames.

---

Frequently Asked Questions

Is Computer Networks difficult for BSc CSIT students? CN is one of the more concept-heavy subjects in the CSIT curriculum because it combines mathematical problems (subnetting, Dijkstra), protocol understanding (TCP, DNS), and theoretical models (OSI). Students who spread their preparation across the full semester find it manageable; those who try to cram it in the final week typically struggle with subnetting and routing.

Which unit is most important for the exam? Unit 4 (Network Layer) carries 10 teaching hours and almost always generates the most exam marks — subnetting and routing algorithms together account for a large portion of long-question marks. OSI (Unit 1) and TCP (Unit 5) follow closely.

Does subnetting come every year? Based on consistent past paper patterns, yes — subnetting appears in almost every TU CN exam, either as a standalone numerical question or embedded in a broader IP addressing question.

Which routing algorithm is asked most? Dijkstra's algorithm appears more frequently than Bellman-Ford in TU exams, typically as a step-by-step worked example with a provided graph. Practice building the distance table correctly.

What's the difference between TCP and UDP in one sentence? TCP is connection-oriented with guaranteed, ordered delivery and flow/congestion control; UDP is connectionless with no guarantees but much lower overhead.

Can I pass CN by studying only past questions? Past questions are a good guide for which topics to prioritize, but TU varies the specific examples (different subnetting prefix, different routing graph). Understanding the process is more reliable than memorizing specific question solutions.

Why does DNS need multiple server types (root, TLD, authoritative)? No single server could store every hostname-to-IP mapping in the world and respond quickly to billions of queries. The hierarchical system distributes the load: root servers know who manages .com, .org, etc.; TLD servers know who manages google.com, facebook.com, etc.; authoritative servers know the actual IP for specific hostnames.

What is the difference between IMAP and POP3? IMAP leaves emails on the server and syncs across all devices; POP3 downloads and removes emails from the server. IMAP is what modern email clients use because it supports multiple devices.

Is the lab component difficult? The CN lab is practical and requires hands-on work in Packet Tracer and Wireshark. Students who actually configure routers, set up VLANs, and analyze packet captures in the lab find the theory easier to understand, so don't treat lab and theory as separate preparation tracks.

What is NAT and why does it exist? NAT (Network Address Translation) maps multiple private IP addresses to a single public IP address. It exists because IPv4 addresses ran out — billions of devices share a much smaller pool of public IPs by hiding behind NAT routers. IPv6 was designed to eliminate the need for NAT by providing enough addresses for every device on earth.

---

Conclusion

Computer Networks is one of those subjects where the exam rewards students who actually understand how the internet works, not just those who've memorized which OSI layer does what. The subnetting math gets fast with practice. Dijkstra's algorithm becomes mechanical once you've done it with three or four different graphs. The protocol comparisons (TCP vs UDP, OSI vs TCP/IP, Leaky vs Token Bucket) make sense once you understand why each design choice was made rather than just what the difference is.

For related guides, see: BSc CSIT Mathematics-I (MTH117) Complete Guide, C Programming (CSC115) Complete Guide, Digital Logic (CSC116) Complete Guide, Introduction to Information Technology (CSC114) Complete Guide.