Start typing to search across all sections...
📚 Content
📋 Syllabus 📄 PYQ Analysis 📊 Unit Trends 🔷 Diagrams 🔢 Numericals
❓ Q&A
📝 Long Questions ⚡ Short Questions 📖 Definitions ⚖️ Comparisons 🃏 Flashcards
🎯 Strategy
🔥 Leaked Paper 🎯 80/20 Strategy ✅ Checklist 🔮 Predictions 📌 Cheat Sheet ⚠️ Exam Traps
BCS603 · AKTU B.TECH SEM VI · 2025–26

Computer
Networks

Aggressive exam-oriented analysis of 8 years of PYQs + leaked rescheduled paper. Every topic quantified, every pattern exposed. Beat the exam in 3 hours.
8
Papers Analyzed
5
Total Units
70
Max Marks
3
Hours
14
Sec A (2m×7)
21
Sec B (7m×3)
35
Sec C (7m×5)
📊 Preparation Tracker 0%
Unit I — Physical Layer0%
Unit II — Data Link0%
Unit III — Network0%
Unit IV — Transport0%
Unit V — Application0%
💡 Click any bar to cycle through 0 → 25 → 50 → 75 → 100%
01

Syllabus Breakdown

5 UNITS
I
Introductory Concepts + Physical Layer
8 Lectures · CO1 (K1, K2)
🔥 CRITICAL
Importance Score
9.5
PYQ Appearances
8/8
Expected Marks
~16m
Difficulty
Med
OSI Model ⭐⭐⭐TCP/IP Model ⭐⭐⭐Network Topologies ⭐⭐⭐Transmission MediaEncoding SchemesSignal/BandwidthSwitching TechniquesMultiplexingShannon TheoremNetwork Devices
II
Data Link Layer
8 Lectures · CO2 (K3)
🔥 CRITICAL
Importance Score
9.5
PYQ Appearances
8/8
Expected Marks
~14m
Difficulty
High
CRC ⭐⭐⭐Hamming Code ⭐⭐FramingPiggybacking ⭐⭐Sliding Window ⭐⭐⭐Go-Back-NSelective RepeatALOHA / Slotted ALOHA ⭐⭐CSMA/CD ⭐⭐CSMA/CALAN Standards
III
Network Layer
8 Lectures · CO3 (K2, K3)
🔥 CRITICAL
Importance Score
9.5
PYQ Appearances
8/8
Expected Marks
~14m
Difficulty
High
IP Addressing ⭐⭐⭐Subnetting ⭐⭐⭐DVR ⭐⭐⭐Link State Routing ⭐⭐Dijkstra's AlgoARP/RARP/DHCP ⭐⭐ICMPIPv4 Header ⭐⭐IPv6 Features ⭐⭐Count-to-Infinity ⭐⭐Congestion Control
IV
Transport Layer
8 Lectures · CO4 (K2, K3)
⭐ HIGH
Importance Score
8.5
PYQ Appearances
7/8
Expected Marks
~14m
Difficulty
Med
TCP vs UDP ⭐⭐⭐TCP Header ⭐⭐3-Way Handshake ⭐⭐Congestion Control ⭐⭐Flow ControlLeaky/Token Bucket ⭐⭐Silly Window SyndromeQoS ⭐⭐
V
Application Layer
8 Lectures · CO5, CO6 (K2)
⭐ HIGH
Importance Score
8.5
PYQ Appearances
7/8
Expected Marks
~14m
Difficulty
Low
DNS ⭐⭐⭐HTTP/HTTPS ⭐⭐FTP ⭐⭐⭐SMTP/IMAP/POP3 ⭐⭐SNMP ⭐⭐⭐TELNET ⭐⭐RSA/Cryptography ⭐⭐VPN
02

PYQ Topic Frequency

8 PAPERS
ℹ️ Based on papers: 2016-17, 2017-18, 2018-19, 2021-22, 2022-23, 2023-24, 2024-25, 2025-26 (rescheduled). Sorted by frequency descending.
#TopicUnitTimes AskedLast AskedPriorityFrequency
1OSI Model — Layers, Functions, DiagramI8/82026🔥 MUST
2Network Topologies (Bus, Star, Ring, Mesh, Tree)I8/82026🔥 MUST
3CRC Error Detection (Numerical + Theory)II7/82026🔥 MUST
4Routing Algorithms (DVR, Dijkstra, Link State)III7/82026🔥 MUST
5DNS — Architecture, Working, RecordsV7/82024🔥 MUST
6TCP Header + Features + SegmentsIV7/82024🔥 MUST
7FTP Protocol — Active vs PassiveV7/82024🔥 MUST
8TCP/IP vs OSI Model ComparisonI7/82026🔥 MUST
9Subnetting Numerical (VLSM/CIDR)III6/82024⭐ HIGH
10TCP vs UDP ComparisonIV6/82024⭐ HIGH
11SNMP ProtocolV6/82024⭐ HIGH
12Congestion Control TechniquesIII/IV6/82026⭐ HIGH
13Sliding Window / Go-Back-N / Selective RepeatII6/82026⭐ HIGH
14SMTP / POP3 / IMAP Email ProtocolsV6/82024⭐ HIGH
15RSA Algorithm / Cryptography BasicsV5/82024⭐ HIGH
16IPv4 vs IPv6 Features & ComparisonIII5/82024⭐ HIGH
17CSMA/CD and CSMA/CAII5/82026⭐ HIGH
18Count-to-Infinity ProblemIII5/82024⭐ HIGH
19Slotted ALOHA (Numerical + Theory)II5/82024◆ MED
20QoS ParametersIV5/82023◆ MED
21ARP / RARP / DHCPIII4/82024◆ MED
22IPv4 Header FormatIII4/82024◆ MED
23Three-Way Handshake (TCP)IV4/82024◆ MED
24PiggybackingII4/82024◆ MED
25TELNET ProtocolV4/82024◆ MED
26HTTP / HTTPSV4/82024◆ MED
27Transmission/Propagation Delay NumericalsI/II4/82026◆ MED
28Hamming Code (Numerical)II3/82023◆ MED
29Token Bucket / Leaky BucketIV3/82024✓ LOW
30Shannon Capacity / Nyquist TheoremI3/82023✓ LOW
03

Unit-wise Trend Analysis

UNIT I Introductory Concepts + Physical Layer 100% Papers ▼
🔁 MOST REPEATED
  • ✦ OSI Model (all 8 papers)
  • ✦ Network Topologies (all 8 papers)
  • ✦ TCP/IP vs OSI comparison (7/8)
  • ✦ Network Devices (Hub/Switch/Router)
  • ✦ Transmission Media (Twisted, Coaxial, Fiber)
📝 THEORY QUESTIONS
  • Applications of Computer Networks
  • Advantages/disadvantages of each topology
  • Functions of each OSI layer with protocols
  • Layering principles & services
  • Encoding: NRZ-L, NRZ-I, Manchester
🔢 NUMERICALS
  • Mesh topology links: n(n-1)/2
  • Shannon capacity: B·log₂(1+S/N)
  • Nyquist rate: 2·B·log₂(M)
  • Propagation delay = distance/speed
  • Transmission time = size/bandwidth
🎨 DIAGRAMS TO DRAW
  • OSI 7-layer model with protocols
  • TCP/IP 4-layer model
  • 5 topology diagrams
  • NRZ-L, NRZ-I, Manchester encoding
UNIT II Data Link Layer 100% Papers ▼
🔁 MOST REPEATED
  • ✦ CRC Error Detection (7/8 papers)
  • ✦ Sliding Window — GBN, SR (6/8)
  • ✦ ALOHA / Slotted ALOHA (5/8)
  • ✦ CSMA/CD, CSMA/CA (5/8)
  • ✦ Piggybacking (4/8 — 3 consecutive)
🔢 KEY NUMERICALS
  • CRC polynomial division
  • Slotted ALOHA max throughput = 1/e ≈ 36.8%
  • Stop-and-Wait efficiency = 1/(1+2a)
  • GBN window = 2ⁿ−1, SR window = 2ⁿ⁻¹
  • Hamming code bit positions
⚠️ HOT TIPS
Piggybacking appeared in 2022-23, 2023-24, 2024-25 — 3 consecutive years. Must know the concept + advantages. Stop-and-Wait disadvantages appeared in 2021-22.
UNIT III Network Layer 100% Papers ▼
🔁 MOST REPEATED
  • ✦ Routing Algorithms (7/8)
  • ✦ Subnetting Numerical (6/8)
  • ✦ Count-to-Infinity (5/8)
  • ✦ IPv4 vs IPv6 (5/8)
  • ✦ IPv4 Header Format (4/8)
🔢 KEY NUMERICALS
  • DVR — Bellman-Ford distance tables
  • Dijkstra's — shortest path step-by-step
  • Subnetting — hosts, network ID, broadcast
  • CIDR notation conversion
⚠️ HOT TIPS
DVR + Count-to-Infinity are always paired in questions. Learn the split-horizon and poison-reverse solutions. Subnetting numerical appears in almost every exam — practice 3-4 examples.
UNIT IV Transport Layer 88% Papers ▼
🔁 MOST REPEATED
  • ✦ TCP vs UDP (6/8)
  • ✦ TCP Header Format (7/8)
  • ✦ Congestion Control (6/8)
  • ✦ Three-Way Handshake (4/8)
  • ✦ QoS Parameters (5/8)
📝 THEORY FOCUS
  • Flow control vs Congestion control
  • Leaky Bucket vs Token Bucket algorithm
  • Silly Window Syndrome + Nagle's Algo
  • Slow start, Congestion avoidance, Fast retransmit
UNIT V Application Layer 88% Papers ▼
🔁 MOST REPEATED
  • ✦ DNS (7/8)
  • ✦ FTP — Active vs Passive (7/8)
  • ✦ SNMP (6/8)
  • ✦ SMTP / Email Protocols (6/8)
  • ✦ RSA/Cryptography (5/8)
📝 THEORY FOCUS
  • DNS hierarchy — root, TLD, authoritative
  • Email flow: MUA → MTA → MDA
  • SNMP components: Manager, Agent, MIB
  • RSA key generation steps + small numerical
04

Must-Know Diagrams

EXAM READY
⚡ These ASCII diagrams mirror exactly what you must draw in the exam. Copy the structure, label every field. Draw first, explain second. 
🔷 OSI 7-Layer Model8/8 papers 🔥 MUST
+---+-------------------+-------------------------------+------------------+ | # | LAYER NAME | KEY PROTOCOLS | PDU NAME | +---+-------------------+-------------------------------+------------------+ | 7 | Application | HTTP, FTP, SMTP, DNS, SNMP | Data / Message | | 6 | Presentation | SSL/TLS, JPEG, ASCII, MPEG | Data | | 5 | Session | NetBIOS, RPC, NFS, SIP | Data | | 4 | Transport | TCP, UDP | Segment/Datagram | | 3 | Network | IP, ICMP, ARP, OSPF, RIP | Packet | | 2 | Data Link | Ethernet, 802.11, PPP, HDLC | Frame | | 1 | Physical | RS-232, DSL, 802.3 physical | Bits | +---+-------------------+-------------------------------+------------------+ TCP/IP Model (4 layers): Application | Transport | Internet | Network Access Mnemonic (top-down): "All People Seem To Need Data Processing"
📦 TCP Header Format (minimum 20 bytes)7/8 papers 🔥 MUST
0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Source Port (16 bits) | Destination Port (16 bits)| -- bytes 0-3 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Sequence Number (32 bits) | -- bytes 4-7 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Acknowledgment Number (32 bits) | -- bytes 8-11 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |HL |Rsvd | U A P R S F | Window Size (16 bits) | -- bytes 12-15 |(4)|(6b) | R C S S Y I | | | | | G K H T N N | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Checksum (16 bits) | Urgent Pointer (16 bits) | -- bytes 16-19 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Options (variable, padded to 32 bits) | -- bytes 20+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ Flags: URG ACK PSH RST SYN FIN | Min header = 5x32bits = 20 bytes
📊 IPv4 Header Format (minimum 20 bytes)4/8 papers ⭐ HIGH
0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | VER | IHL | DSCP (6) |ECN| Total Length (16) | | (4b) | (4b) | |(2)| | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Identification (16 bits)| Flags| Fragment Offset (13) | | | (3b) | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | TTL (8) | Protocol (8) | Header Checksum (16) | | | 6=TCP 17=UDP | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Source IP Address (32 bits) | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Destination IP Address (32 bits) | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Options + Padding (variable, optional) | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ Flags: bit0=reserved | DF=Do Not Fragment | MF=More Fragments coming
🔄 TCP 3-Way Handshake + 4-Way Close4/8 papers ⭐ HIGH
CLIENT SERVER | | |------- SYN (seq=x) ---------------------------> | STEP 1: Client initiates | | |<------ SYN-ACK (seq=y, ack=x+1) -------------- | STEP 2: Server responds | | |------- ACK (seq=x+1, ack=y+1) ----------------> | STEP 3: Client confirms | | |======= CONNECTION ESTABLISHED ================= | | | |------- DATA -----------------------------------> | Data transfer begins |<------ ACK ------------------------------------ | | | 4-WAY CLOSE (either side can initiate): |------- FIN ------------------------------------> | Half-close: I am done sending |<------ ACK ------------------------------------ | Server acknowledges FIN |<------ FIN ------------------------------------ | Server done sending too |------- ACK ------------------------------------> | Client acknowledges | (TIME_WAIT: client waits 2*MSL ~2 min) |
📬 DNS Hierarchy + Resolution Process7/8 papers 🔥 MUST
+------------------+ | ROOT SERVERS | (.) 13 root server clusters globally +------------------+ / | \ +-------+ +------+ +------+ | .com | | .org | | .in | TLD Servers (Top-Level Domain) +-------+ +------+ +------+ | +----------+ | google | Authoritative Nameserver (Second-Level Domain) | .com | Has the actual A/AAAA/MX records +----------+ | +----------+ | www | Host record - returns IP address +----------+ RECURSIVE: Client --(1)--> Local DNS --(2)(3)(4)--> returns IP to client ITERATIVE: Client --(1)--> Local DNS (referral) --(2)--> TLD --(3)--> Auth DNS Record Types: A(IPv4) AAAA(IPv6) MX(mail) CNAME(alias) NS(nameserver) PTR(reverse)
🏃 GBN vs SR Sliding Window Timeline6/8 papers 🔥 HIGH
GO-BACK-N (n=3, W=7): SELECTIVE REPEAT (n=3, W=4): Sender: 0 1 2 3* 4 5 6 Sender: 0 1 2 3* 4 5... [ERROR at frame 3!] [ERROR at frame 3!] GBN Response: SR Response: Receiver discards 4,5,6 Receiver BUFFERS 4,5 (already received) Sender retransmits 3 4 5 6 0... Sender retransmits 3 ONLY Wasteful but simple receiver Efficient but complex receiver Sender window: W <= 2^n - 1 = 7 Sender window: W <= 2^(n-1) = 4 Receiver window: 1 Receiver window: 4 (needs buffer) Efficiency(W<1+2a): W/(1+2a) Efficiency(W<1+2a): W/(1+2a) KEY: SR window is half of GBN to prevent old/new frame ambiguity at receiver!
05

Worked Numericals

STEP-BY-STEP
✅ These are the exact numerical types that appear in AKTU CN exam. Practice each type at least twice.
01
CRC Error Detection
Unit II · Asked 7/8 papers · 🔥 MOST IMPORTANT
7m TYPE
Q: The data word is 1101011011. The generator polynomial is x⁴+x+1 (10011). Find the CRC code sent. Verify if the received codeword has any error.
  1. STEP 1
    Append (n-1) zeros to data: Generator degree = 4, so append 4 zeros.
    Data with zeros: 11010110110000
  2. STEP 2
    Perform binary (XOR) division of data+zeros by generator 10011:
    11010110110000 ÷ 10011
    XOR 10011 with 11010 → 01001
    Bring down → 10011 → XOR 10011 → 00000
    Bring down → 01101 → XOR 00000 → continue...
    Continue until 4 bits remain as remainder.
    Remainder (CRC) = 1110
  3. STEP 3
    Transmitted codeword = Data + CRC
    1101011011 + 1110 = 11010110111110
  4. STEP 4
    Verification: Receiver divides received codeword by same generator. If remainder = 0000 → No error. If non-zero → Error detected.
    ✓ If remainder = 0000 → Frame accepted (no error)
📌 Key formula: CRC bits = degree of generator polynomial. Generator always starts and ends with 1.
02
Subnetting — Network ID, Broadcast, Host Range
Unit III · Asked 6/8 papers · ⭐ HIGH PRIORITY
7m TYPE
Q: An organization has been assigned the IP address 192.168.10.0/26. Find the subnet mask, number of hosts, network ID, broadcast address, and valid host range.
  1. STEP 1
    Subnet Mask: /26 means 26 bits are 1s.
    11111111.11111111.11111111.11000000
    Subnet Mask = 255.255.255.192
  2. STEP 2
    Host bits = 32 − 26 = 6
    Total addresses = 2⁶ = 64
    Usable hosts = 64 − 2 = 62 (subtract Network ID and Broadcast)
  3. STEP 3
    Network ID = Host bits all 0s → 192.168.10.0
    Broadcast = Host bits all 1s → 192.168.10.63
    First host = 192.168.10.1
    Last host = 192.168.10.62
📌 Formula: Hosts = 2^(host bits) − 2. Classes: A = /8, B = /16, C = /24.
03
Distance Vector Routing (Bellman-Ford)
Unit III · Asked 7/8 papers
7m TYPE
Q: Consider routers A, B, C with costs: A-B=2, A-C=6, B-C=3. Construct initial routing tables and show DVR exchange for 1 iteration.
  1. STEP 1
    Initial tables (each router knows only directly connected costs):
    A: [A=0, B=2, C=6] | B: [A=2, B=0, C=3] | C: [A=6, B=3, C=0]
  2. STEP 2
    A updates using B's table: Cost via B to C = cost(A,B) + cost(B,C) = 2+3 = 5 < 6
    A updates: C = 5 via B (was 6 direct)
  3. STEP 3
    Bellman-Ford: D(x,y) = min over all neighbors n { c(x,n) + D(n,y) }
    Converged table: A→B=2, A→C=5(via B)
04
ALOHA Throughput Calculation
Unit II · Asked 5/8 papers
2-7m TYPE
Q: Find the maximum throughput of (a) Pure ALOHA and (b) Slotted ALOHA. If frame generation rate G = 0.5, find actual throughput for both.
  1. FORMULAS
    Pure ALOHA: S = G · e−2G (max at G=0.5 → S=1/2e ≈ 18.4%)
    Slotted ALOHA: S = G · e−G (max at G=1 → S=1/e ≈ 36.8%)
  2. CALC
    At G = 0.5:
    Pure: S = 0.5 × e−1 = 0.5 × 0.368 = 0.184 = 18.4%
    Slotted: S = 0.5 × e−0.5 = 0.5 × 0.607 = 0.303 = 30.3%
05
Shannon Capacity + Nyquist Theorem
Unit I · Appeared 3/8 papers
2-7m TYPE
Q: A telephone channel has bandwidth 3000 Hz and SNR = 30 dB. Find (a) Shannon capacity and (b) Nyquist rate if 8 signal levels are used.
  1. SHANNON
    SNR(dB) = 30 → SNR = 10^(30/10) = 1000
    C = B · log₂(1 + SNR) = 3000 × log₂(1001)
    C = 3000 × 9.97 ≈ 29,910 bps ≈ 30 kbps
  2. NYQUIST
    N = 2 × B × log₂(M) where M = 8 signal levels
    N = 2 × 3000 × 3 = 18,000 bps = 18 kbps
06
Stop-and-Wait / Sliding Window Efficiency
Unit II · Appeared 4/8 papers
7m TYPE
Q: Link bandwidth = 1 Mbps, frame size = 1000 bits, propagation delay = 10 ms. Find efficiency of Stop-and-Wait, GBN (window=7), and SR (window=4).
  1. STEP 1
    Transmission time Tₜ = L/B = 1000/10⁶ = 1 ms
    Propagation delay Tₚ = 10 ms
    a = Tₚ/Tₜ = 10/1 = 10
  2. STEP 2
    Stop-and-Wait: η = 1/(1+2a) = 1/(1+20) = 1/21 ≈ 4.76%
  3. STEP 3
    GBN (W=7): Since W=7 < 1+2a=21: η = W/(1+2a) = 7/21 = 33.3%
    SR (W=4): η = 4/21 = 19%
06

Long Questions (7-mark)

Sec B+C
⚡ Click "Show Answer Outline" to see model answer structure for each question.
Q01 · Unit I · 8/8 papers
Explain the OSI reference model. Describe the functions of each layer with examples of protocols used at each layer. Draw the diagram.
🔥 MUST PREPARE7 marksDiagram required
  • Introduction: OSI = Open Systems Interconnection, 7-layer model by ISO, provides framework for protocol design.
  • Physical (L1): Bit transmission, electrical signals, cables, hubs. Protocols: RS-232, DSL.
  • Data Link (L2): Framing, error detection, MAC addressing. Protocols: Ethernet (802.3), Wi-Fi (802.11).
  • Network (L3): Routing, logical addressing, IP. Protocols: IP, ICMP, ARP.
  • Transport (L4): End-to-end delivery, flow control, error recovery. Protocols: TCP, UDP.
  • Session (L5): Establish/maintain/terminate sessions. Protocols: NetBIOS, RPC.
  • Presentation (L6): Data translation, encryption, compression. Protocols: SSL/TLS, JPEG.
  • Application (L7): User interface, network services. Protocols: HTTP, FTP, SMTP, DNS.
  • Draw: 7 stacked boxes with layer names + protocol examples + encapsulation arrow.
Q02 · Unit II · 7/8 papers
Explain CRC error detection method. Given data word D = 10110011 and generator G = x³+x+1, find the CRC and the transmitted frame. Verify error detection.
🔥 MUST PREPARE7 marksNumerical
  • CRC Concept: Cyclic Redundancy Check — polynomial division. Sender appends r bits (degree of generator), receiver divides, if remainder=0 no error.
  • Generator x³+x+1 = 1011 (4 bits, degree 3)
  • Step 1: Append 3 zeros → 10110011000
  • Step 2: XOR divide 10110011000 ÷ 1011 step by step
  • Step 3: Get remainder (CRC bits)
  • Step 4: Transmitted frame = Data + CRC
  • Verification: Divide received frame by generator; remainder=0 means no error
  • Properties of CRC: Detects all burst errors up to r bits, all odd errors, all double-bit errors.
Q03 · Unit III · 7/8 papers
Explain Distance Vector Routing algorithm. Construct routing tables and show one complete exchange. What is the Count-to-Infinity problem and how is it solved?
🔥 MUST PREPARE7 marksTable numerical
  • DVR Concept: Based on Bellman-Ford. Each router maintains table with distance to all routers. Shares table with neighbors periodically.
  • Algorithm: D(x,y) = min{c(x,n) + D(n,y)} for all neighbors n
  • Iteration: Show initial tables → one exchange → updated tables (use 3-node triangle example)
  • Count-to-Infinity: When a link fails, bad info propagates slowly (increments by 1 each exchange), distance grows to infinity before realizing route is dead.
  • Solutions: (a) Split Horizon: don't advertise route back to where you learned it. (b) Poison Reverse: advertise route with metric=∞. (c) Hold-down timers. (d) Route poisoning.
  • Advantages: Simple, low overhead. Disadvantages: Slow convergence, count-to-infinity.
Q04 · Unit V · 7/8 papers
Explain the Domain Name System (DNS). Draw the DNS hierarchy. Explain recursive and iterative name resolution with an example.
🔥 MUST PREPARE7 marksDiagram required
  • DNS Purpose: Maps human-readable domain names to IP addresses. Distributed, hierarchical database.
  • Hierarchy: Root (.) → TLD (.com, .org, .in, .edu) → Second-Level (google.com) → Subdomain (www.google.com)
  • Record Types: A (IPv4 address), AAAA (IPv6), MX (mail server), CNAME (alias), NS (name server), PTR (reverse lookup)
  • Recursive: Client asks local DNS → local DNS does all work, returns final answer. One query from client.
  • Iterative: Client asks local DNS → local DNS asks root → root refers to TLD → TLD refers to authoritative → local DNS returns IP. Client gets intermediate referrals.
  • Caching: DNS responses cached with TTL to reduce load.
  • Draw: Tree diagram + resolution arrow diagram (recursive vs iterative)
Q05 · Unit IV · 6/8 papers
Compare TCP and UDP in detail. Explain the TCP three-way handshake. When would you choose UDP over TCP?
⭐ HIGH7 marksComparison table
  • TCP: Connection-oriented, reliable, ordered, flow control, congestion control, full-duplex, heavyweight header (20+ bytes). Used for: HTTP, FTP, SMTP.
  • UDP: Connectionless, unreliable, no ordering, no flow/congestion control, lightweight (8-byte header). Used for: DNS, VoIP, streaming, gaming.
  • Three-Way Handshake: SYN(x) → SYN-ACK(y, ack=x+1) → ACK(ack=y+1). Establishes seq numbers for both directions.
  • Choose UDP when: Speed > reliability, app handles errors itself, broadcast/multicast needed, real-time applications.
  • Draw: Comparison table (10 rows) + handshake timeline diagram
Q06 · Unit II · 6/8 papers
Explain Go-Back-N and Selective Repeat ARQ protocols. Derive window size formulas. Compare their performance with Stop-and-Wait.
⭐ HIGH7 marks
  • Stop-and-Wait: Send one frame, wait for ACK. Window=1. Efficiency = 1/(1+2a).
  • Go-Back-N: Window size W=2ⁿ−1. On error, retransmit all unacknowledged frames from error. Receiver discards all out-of-order frames. Simple receiver but wasteful.
  • Selective Repeat: Window size W=2ⁿ⁻¹. Only retransmit erroneous frame. Receiver buffers out-of-order frames. Complex receiver but efficient.
  • Efficiency: If W ≥ 1+2a: η=1 (100%). If W < 1+2a: η=W/(1+2a) for GBN, η=W/(1+2a) for SR.
  • Why 2ⁿ⁻¹ for SR? To distinguish old and new frames at receiver — prevents ambiguity.
Q07 · Unit V · 6/8 papers
Explain SNMP (Simple Network Management Protocol). Describe its components, versions, and operations (Get, GetNext, Set, Trap).
⭐ HIGH7 marks
  • SNMP: Application-layer protocol for network management. Uses UDP port 161 (agent) and 162 (manager/trap).
  • Components: (1) Manager (NMS) — sends queries. (2) Agent — runs on devices, responds to queries. (3) MIB (Management Information Base) — database of managed objects.
  • Operations: GET (retrieve value), GET-NEXT (next in MIB), SET (change value), TRAP (agent notifies manager of event — unsolicited).
  • Versions: SNMPv1 (simple, community strings), SNMPv2c (bulk operations, improved error handling), SNMPv3 (authentication, encryption, security).
  • MIB Structure: Hierarchical tree of OIDs (Object Identifiers). Enterprise-specific MIBs possible.
  • Draw: Manager ↔ Agent with MIB, show GET/TRAP arrows.
Q08 · Unit V · 5/8 papers
Explain RSA public-key cryptography algorithm with a numerical example. Generate keys, encrypt and decrypt a message.
⭐ HIGH7 marksNumerical
  • Key Generation: Choose p=3, q=11. n=p×q=33. φ(n)=(p-1)(q-1)=20. Choose e coprime to 20 → e=7. Find d: e×d ≡ 1 mod φ(n) → d=3 (since 7×3=21≡1 mod 20).
  • Public key: (e,n) = (7,33). Private key: (d,n) = (3,33).
  • Encrypt: C = Mᵉ mod n. If M=2: C = 2⁷ mod 33 = 128 mod 33 = 29.
  • Decrypt: M = Cᵈ mod n = 29³ mod 33 = 24389 mod 33 = 2 ✓
  • Security: Based on difficulty of factoring large n=p×q. RSA is asymmetric (different keys for encryption/decryption).
Q09 · Unit II · 5/8 papers
Explain Go-Back-N and Selective Repeat ARQ protocols. Derive formulas for window size. Also explain Stop-and-Wait and compare efficiency of all three protocols with a numerical example.
⭐ HIGH7 marksNumerical
  • Stop-and-Wait: Sender transmits one frame and waits for ACK before sending next. Window=1. Efficiency η = 1/(1+2a) where a = Tp/Tt. Wasted time = 2Tp.
  • Go-Back-N: Sender can send up to W = 2ⁿ−1 unacknowledged frames. On error, retransmit ALL frames from the error onward. Receiver discards out-of-order frames. Simple receiver buffer (size=1).
  • Selective Repeat: Only retransmit the specific errored frame. Window W = 2^(n−1). Receiver buffers out-of-order frames. Efficient but complex receiver.
  • Efficiency (W ≥ 1+2a): η = 100%. If W < 1+2a: GBN η = W/(1+2a), SR η = W/(1+2a).
  • Numerical: BW=1Mbps, L=1000b, Tp=10ms → Tt=1ms, a=10. S&W: 1/21=4.76%. GBN(W=7): 7/21=33.3%. SR(W=4): 4/21=19%.
  • Why W≤2^(n-1) for SR? To prevent receiver from confusing old frames with new ones (ambiguity in sequence number space).
Q10 · Unit I · 8/8 papers
Compare and contrast the OSI and TCP/IP reference models. What are the advantages and disadvantages of layered architectures? Explain the concept of encapsulation.
🔥 MUST7 marksDiagram
  • OSI: 7 layers — Physical, Data Link, Network, Transport, Session, Presentation, Application. By ISO. Theoretical model. Protocol-independent. Each layer has specific services.
  • TCP/IP: 4 layers — Network Access (Physical+DLL), Internet (Network), Transport, Application. By DARPA. Practical, used in actual Internet. Protocol-specific.
  • Key Differences: OSI=7 layers, TCP/IP=4; OSI came before protocols, TCP/IP after; OSI has Session+Presentation, TCP/IP merges them into Application.
  • Advantages of Layering: Modularity, easier debugging, plug-and-play (change one layer without affecting others), standardization, abstraction.
  • Disadvantages: Overhead from multiple headers, sometimes layers duplicate services (error control appears at multiple layers), rigidity.
  • Encapsulation: Each layer adds its own header (and trailer at DLL) to the data from upper layer. Data→Segment→Packet→Frame→Bits as we go down. Decapsulation at receiver removes headers layer by layer.
Q11 · Unit III · 5/8 papers
Explain IPv4 header format with a neat diagram. Describe the significance of each field. What is fragmentation and how is it handled by the IP header?
⭐ HIGH7 marksHeader diagram
  • IPv4 Header Fields (20 bytes minimum):
  • VER(4b): IP version (4). IHL(4b): Header length in 32-bit words (min=5). DSCP(8b): Differentiated services for QoS. Total Length(16b): Header+data in bytes.
  • ID(16b): Unique fragment identification. Flags(3b): DF (don't fragment), MF (more fragments). Fragment Offset(13b): Position of fragment in original datagram.
  • TTL(8b): Hop limit, decremented by each router; dropped at 0. Protocol(8b): Upper layer protocol (6=TCP, 17=UDP, 1=ICMP). Header Checksum(16b): Error detection for header only.
  • Source IP(32b): Sender's IP. Destination IP(32b): Receiver's IP. Options: Variable, for testing/debugging.
  • Fragmentation: When datagram > MTU of outgoing link, router breaks it into fragments. Each fragment has same ID, MF bit set (last has MF=0), and Fragment Offset. Reassembly at destination only.
Q12 · Unit II · 4/8 papers
Explain Piggybacking. What are its advantages? How does it differ from a pure ACK-based system? Give examples where piggybacking is used.
◆ MEDIUM7 marks
  • Piggybacking: When a receiver has data to send to the sender, it delays its ACK briefly and attaches it to the outgoing data frame instead of sending a separate ACK packet.
  • How it works: Host A sends data to B → B has data to send back → B attaches ACK (for A's frame) in the header of its own data frame → A receives data + ACK in one frame.
  • Advantages: (1) Reduces number of frames sent → reduces network traffic. (2) Saves bandwidth — one frame serves dual purpose. (3) Reduces processing overhead at both ends. (4) Halves number of frames in full-duplex scenario.
  • Disadvantages: If B doesn't have data to send soon, ACK must eventually be sent separately (delay bounded by timer). Can cause slight ACK delay.
  • Examples: Used in TCP (ACK field in every TCP segment), HDLC protocol, PPP. Any full-duplex protocol can use piggybacking.
  • Comparison: Pure ACK system sends 2×frames (data+separate ACK). Piggybacking sends 1×frame carrying both — doubles efficiency when traffic is bidirectional.
Q13 · Unit IV · 4/8 papers
What is congestion in a network? Explain the congestion control mechanisms used by TCP: Slow Start, Congestion Avoidance, Fast Retransmit, and Fast Recovery.
⭐ HIGH7 marks
  • Congestion: Occurs when too many packets are in the network, causing queues to fill, packets dropped, increased delay, and reduced throughput. Like a traffic jam.
  • TCP Congestion Window (CWND): Limits how many bytes sender can have in-flight. Effective window = min(CWND, receiver's advertised window).
  • Slow Start: CWND starts at 1 MSS. Doubles every RTT (exponential growth) until it reaches ssthresh (slow start threshold). Goal: probe network capacity quickly but safely.
  • Congestion Avoidance: Once CWND ≥ ssthresh, increase CWND by 1 MSS per RTT (linear/additive increase). This is AIMD — Additive Increase Multiplicative Decrease.
  • Fast Retransmit: If 3 duplicate ACKs received for same segment → assume that segment lost (not timeout). Retransmit immediately without waiting for timeout. More efficient than waiting for timeout.
  • Fast Recovery: After fast retransmit, set ssthresh = CWND/2, CWND = ssthresh + 3 (already received 3 duplicates). Continue with congestion avoidance (not slow start). Prevents fall to CWND=1.
  • On Timeout: ssthresh = CWND/2, CWND = 1 MSS, restart slow start. Timeout = full packet loss, more severe than triple duplicate ACK.
Q14 · Unit V · 4/8 papers
Explain FTP (File Transfer Protocol) in detail. What is the difference between Active and Passive FTP? Draw the connection establishment diagram for both modes.
🔥 MUST7 marksDiagram
  • FTP: Application layer protocol for transferring files between client and server. Uses TWO separate TCP connections: Control connection (port 21) and Data connection (port 20 or dynamic).
  • Control Connection: Always initiated by client to server port 21. Used for commands (USER, PASS, LIST, RETR, STOR) and responses. Persistent throughout session.
  • Active FTP: (1) Client connects to server port 21 (control). (2) Client sends PORT command with its IP and port number. (3) SERVER initiates data connection from port 20 → client's specified port. Problem: Client's firewall may block incoming connection from server.
  • Passive FTP: (1) Client connects to server port 21 (control). (2) Client sends PASV command. (3) Server responds with its IP and random port. (4) CLIENT initiates data connection to server's port. Firewall-friendly — all connections initiated by client.
  • FTP Commands: USER, PASS (authentication), LIST (directory listing), RETR (download), STOR (upload), QUIT (logout), CWD (change directory).
  • Advantages: Reliable transfer (TCP), resume support, authentication. Disadvantages: Plaintext (no encryption) — use SFTP or FTPS for security. Two connections add complexity.
Q15 · Unit II · 5/8 papers
Explain CSMA/CD and CSMA/CA protocols. How does exponential backoff work? Compare these protocols and state when each is used.
⭐ HIGH7 marks
  • CSMA/CD (Collision Detection): Used in wired Ethernet (IEEE 802.3). Steps: (1) Listen before transmitting (Carrier Sense). (2) If medium busy, wait. (3) Transmit. (4) If collision detected, stop and send 32-bit jam signal. (5) Wait random backoff time, retry.
  • Binary Exponential Backoff: After kth collision, wait random time chosen from {0, 1, ..., 2^k−1} × slot time. Max k=10. After 16 failures, report error. Reduces collision probability geometrically.
  • CSMA/CA (Collision Avoidance): Used in Wi-Fi (IEEE 802.11). Collisions CANNOT be detected while transmitting wirelessly (hidden terminal problem). Strategy: AVOID collisions before they happen.
  • CSMA/CA Process: (1) Sense medium (DIFS wait after idle). (2) If idle, transmit after DIFS. (3) If busy, set random backoff counter. (4) Count down only while medium idle. (5) Transmit when counter=0. (6) Wait for ACK from receiver.
  • RTS/CTS Mechanism: Optional. Sender sends RTS (Request to Send), receiver sends CTS (Clear to Send). All nearby nodes hear CTS and defer. Solves hidden terminal problem.
  • Comparison: CSMA/CD — wired, detects collision, no ACK needed. CSMA/CA — wireless, avoids collision, ACK required for every frame. CD is more efficient; CA has higher overhead.
Q16 · Unit II · 7/8 papers
Explain Hamming code for error correction. Given data bits D = 1011, find the Hamming code. Locate and correct the error if received codeword is 0111001.
🔥 MUST PREPARE7 marksNumerical
  • Concept: Hamming code inserts parity bits (r) at positions 2⁰, 2¹, 2², ... (1, 2, 4, 8...). Condition: 2^r ≥ m + r + 1. For m=4: try r=3 → 2³=8 ≥ 4+3+1=8 ✓. Total bits = 4+3 = 7.
  • Bit positions: 1=P1, 2=P2, 3=D1, 4=P4, 5=D2, 6=D3, 7=D4. Data 1011 → D1=1, D2=0, D3=1, D4=1.
  • P1 (pos 1): Covers bits 1,3,5,7 → P1 ⊕ D1 ⊕ D2 ⊕ D4 = P1 ⊕ 1 ⊕ 0 ⊕ 1 = 0 → P1=0
  • P2 (pos 2): Covers bits 2,3,6,7 → P2 ⊕ D1 ⊕ D3 ⊕ D4 = P2 ⊕ 1 ⊕ 1 ⊕ 1 = 0 → P2=1
  • P4 (pos 4): Covers bits 4,5,6,7 → P4 ⊕ D2 ⊕ D3 ⊕ D4 = P4 ⊕ 0 ⊕ 1 ⊕ 1 = 0 → P4=0
  • Transmitted codeword: Position: 1=0, 2=1, 3=1, 4=0, 5=0, 6=1, 7=1 → 0110011
  • Error detection: Received = 0111001. Compute syndrome: C1 = P1⊕3⊕5⊕7 bits, C2 = P2⊕3⊕6⊕7, C4 = P4⊕5⊕6⊕7. Syndrome = binary position of error → flip that bit → corrected word.
  • Key Rule: Hamming distance = 3 → corrects 1 error, detects 2 errors. Redundancy = r parity bits only.
Q17 · Unit III · 6/8 papers
Explain Link State Routing (LSR). How does it differ from Distance Vector Routing? Apply Dijkstra's algorithm on a given network graph to find the shortest path tree.
🔥 MUST PREPARE7 marksAlgorithm + Table
  • LSR Concept: Each router floods Link State Packets (LSPs) to ALL routers in the network. Every router builds a complete topology map. Then runs Dijkstra's algorithm locally to find shortest paths. Used in OSPF, IS-IS.
  • LSP Contents: Router ID, list of neighbors, link costs, sequence number, TTL.
  • Dijkstra's Algorithm: (1) Set source cost=0, all others=∞. (2) Add source to Confirmed set. (3) For all neighbors of newly confirmed node, compute tentative cost = cost_to_node + link_cost. (4) Pick minimum tentative → add to Confirmed. (5) Repeat until all confirmed.
  • Example (5-node graph): Draw table with columns: Step | Confirmed | Tentative. Fill row by row as nodes get confirmed.
  • DVR vs LSR: DVR = share tables with neighbors only, slow convergence, count-to-infinity. LSR = flood to all, each builds complete map, faster convergence, higher memory/CPU, no loops.
  • LSR Advantages: No count-to-infinity, fast convergence, accurate topology. Disadvantages: High overhead for large networks (flooding), complex implementation.
Q18 · Unit IV · 6/8 papers
Draw and explain the TCP segment header format. What is the significance of sequence number, acknowledgement number, and the control flags? Explain TCP connection termination (4-way handshake).
🔥 MUST PREPARE7 marksHeader diagram
  • TCP Header (min 20 bytes): Source Port (16b), Destination Port (16b), Sequence Number (32b), Acknowledgement Number (32b), Header Length (4b), Reserved (6b), Flags (6b), Window Size (16b), Checksum (16b), Urgent Pointer (16b), Options (variable).
  • Sequence Number: Byte number of first data byte in this segment. Enables ordered reassembly at receiver. Initial seq chosen randomly during SYN (prevents hijacking).
  • ACK Number: Next expected byte from sender. Piggybacks the acknowledgement. Only valid when ACK flag is set.
  • Control Flags: SYN (sync seq during handshake), ACK (ACK field valid), FIN (no more data, close), RST (abort connection), PSH (flush to application immediately), URG (urgent data).
  • Window Size: Advertised by receiver — number of bytes it can buffer. Implements flow control. Can change dynamically.
  • 4-Way Termination: Client→FIN→Server | Server→ACK→Client | Server→FIN→Client | Client→ACK→Server. Client enters TIME_WAIT (2×MSL ≈ 2 min) before fully closing. Full-duplex means each direction closes independently.
Q19 · Unit I · 5/8 papers
Explain the various transmission media used in computer networks. Compare guided and unguided media. Describe twisted pair, coaxial cable, fiber optic, and their characteristics.
⭐ HIGH7 marksComparison table
  • Guided Media (wired): Signal travels through physical medium with defined boundaries.
  • Twisted Pair (UTP/STP): Two insulated copper wires twisted together. Twisting reduces EMI/crosstalk. UTP: up to Cat 6A → 10 Gbps, 100m max. STP: shielded for more noise immunity. Cheap, flexible, most common (Ethernet).
  • Coaxial Cable: Central conductor, insulator, braided shield, outer jacket. 75Ω (cable TV), 50Ω (Ethernet). Better noise immunity than twisted pair. Used in cable TV, older 10Base2 Ethernet.
  • Fiber Optic: Glass/plastic core, cladding (lower refractive index), jacket. Uses total internal reflection to carry light pulses. Single-mode: thin core (~8µm), long distance (km), laser source, expensive. Multi-mode: thick core (~50µm), short distance, LED source, cheaper. Bandwidth: Tbps. Immune to EMI.
  • Unguided Media (wireless): Radio waves (Wi-Fi 2.4/5 GHz, Bluetooth), Microwave (point-to-point, satellite), Infrared (short range, LOS required).
  • Comparison Table: Cost: TP < Coax < Fiber. Distance: TP(100m) < Coax(500m) < Fiber(km). Bandwidth: TP < Coax < Fiber. EMI immunity: TP(low) < Coax(med) < Fiber(immune).
Q20 · Unit III · 6/8 papers
What is subnetting and CIDR (Classless Inter-Domain Routing)? Given a Class C network 192.168.10.0/24, divide it into 4 equal subnets. Find network address, broadcast address, valid host range, and subnet mask for each subnet.
🔥 MUST PREPARE7 marksNumerical
  • Subnetting: Borrowing bits from host portion to create multiple sub-networks. Needed for: efficient IP allocation, security isolation, reduced broadcast domains.
  • To create 4 subnets: Need 2 bits borrowed (2²=4). New mask = /24+2 = /26. Subnet mask = 255.255.255.192. Each subnet has 2⁶=64 addresses, 62 usable hosts.
  • Subnet 1: Network=192.168.10.0, Broadcast=192.168.10.63, Hosts=192.168.10.1–192.168.10.62
  • Subnet 2: Network=192.168.10.64, Broadcast=192.168.10.127, Hosts=192.168.10.65–192.168.10.126
  • Subnet 3: Network=192.168.10.128, Broadcast=192.168.10.191, Hosts=192.168.10.129–192.168.10.190
  • Subnet 4: Network=192.168.10.192, Broadcast=192.168.10.255, Hosts=192.168.10.193–192.168.10.254
  • CIDR: Replaces classful A/B/C with variable-length prefix notation (e.g., /26). Allows route aggregation (supernetting): 4×/24 can be summarized as one /22 in routing tables → reduces routing table size.
Q21 · Unit V · 5/8 papers
Explain the World Wide Web (WWW) and HTTP protocol. Describe HTTP request/response format, HTTP methods, status codes, and the difference between HTTP/1.0, HTTP/1.1, and HTTP/2.
⭐ HIGH7 marksProtocol detail
  • WWW: System of interlinked hypertext documents accessed over the Internet using HTTP. Components: URL (address), HTML (content), HTTP (transfer protocol), Browser (client).
  • HTTP: HyperText Transfer Protocol. Application-layer, stateless, TCP port 80. Client sends Request → Server sends Response.
  • Request Format: Request line (Method URL Version) + Headers + Blank line + Body. Example: GET /index.html HTTP/1.1 | Host: www.example.com
  • HTTP Methods: GET (retrieve resource, no body), POST (send data, has body), PUT (replace resource), DELETE (remove), HEAD (GET without body), OPTIONS (server capabilities).
  • Status Codes: 1xx=Informational, 2xx=Success (200 OK, 201 Created), 3xx=Redirection (301 Moved), 4xx=Client Error (400 Bad Request, 404 Not Found, 403 Forbidden), 5xx=Server Error (500 Internal Server Error).
  • HTTP/1.0: New TCP connection per request (slow). HTTP/1.1: Persistent connections (keep-alive), pipelining, chunked transfer, virtual hosting. HTTP/2: Multiplexing (multiple requests on one connection), header compression (HPACK), server push, binary protocol — much faster.
  • Cookies: Server sends Set-Cookie header → browser stores cookie → browser sends cookie with future requests → enables session state in stateless HTTP.
Q22 · Unit IV · 5/8 papers
Explain Leaky Bucket and Token Bucket algorithms for traffic shaping. Compare the two with a diagram. When is each preferred? Give a numerical example for Token Bucket.
⭐ HIGH7 marksNumerical + Diagram
  • Traffic Shaping: Controlling the rate at which packets are sent into the network to avoid congestion and ensure QoS.
  • Leaky Bucket: Packets enter a bucket (FIFO queue). Bucket drains at constant rate r regardless of input. If bucket full → packet discarded. Output is ALWAYS smooth at rate r — no bursts allowed. Analogy: water drips at constant rate regardless of how fast you pour.
  • Token Bucket: Tokens generated at rate r and stored up to max b. Each packet consumes one token. If tokens available → send packet immediately. If no tokens → packet waits or is dropped. BURSTS allowed up to b packets when tokens accumulated.
  • Numerical: Token bucket: r=2 tokens/sec, b=8 tokens. At t=0, bucket has 8 tokens. Burst of 10 packets arrives → send 8 immediately (tokens exhaust), then 2 per second thereafter. Compare with Leaky Bucket: same 10 packets → 2 per second from start, no burst ever.
  • Key Difference: Leaky Bucket = constant output rate, no burst, smoothing. Token Bucket = can burst up to b, average rate r, more flexible. Token Bucket preferred for applications that generate occasional bursts (video encoding).
  • Diagram: Draw both algorithms side by side with input/output arrows and bucket representation.
Q23 · Unit I · 5/8 papers
Explain multiplexing in computer networks. Describe Frequency Division Multiplexing (FDM), Time Division Multiplexing (TDM), and Statistical TDM with diagrams and comparison.
⭐ HIGH7 marksDiagram required
  • Multiplexing: Technique to combine multiple signals over a single physical link to maximize bandwidth utilization. MUX at sender combines signals; DEMUX at receiver separates them.
  • FDM (Frequency Division): Each channel assigned a unique frequency band with guard bands between them. Signals transmitted simultaneously in different frequencies. Continuous, analog. Used in: radio broadcasting (AM/FM), cable TV, ADSL. Disadvantage: guard bands waste spectrum.
  • TDM (Time Division): Single frequency, time divided into slots. Each channel gets a fixed, repeating time slot in the frame. Synchronous (slots fixed regardless of data) → wastes slots when channel idle. Used in: telephone networks (T1/E1 carriers), SONET/SDH.
  • Statistical TDM (STDM): Slots assigned dynamically — only to channels with data to send. Channel ID included in each slot. More efficient than synchronous TDM — no wasted slots. Used in packet-switched networks, DSL. Disadvantage: variable delay, more complex.
  • WDM (Wavelength Division): Like FDM but for optical fiber — different light wavelengths on same fiber. DWDM supports 100s of channels on one fiber pair → Tbps capacity. Used in long-haul fiber networks.
  • Comparison Table: FDM vs TDM vs STDM: Analog vs Digital, Continuous vs Slots, Fixed vs Dynamic allocation, Guard bands vs Frame overhead.
Q24 · Unit III · 5/8 papers
What is IPv6? Why was it introduced? Explain the IPv6 header format and compare it with IPv4. What are the different types of IPv6 addresses? Explain SLAAC (Stateless Address Autoconfiguration).
⭐ HIGH7 marksHeader comparison
  • Why IPv6: IPv4 address exhaustion (only 4.3 billion addresses). IPv6 = 128-bit address space = 3.4×10³⁸ addresses. Also fixes: no fragmentation by routers, simplified header, built-in IPSec, better multicast, no NAT needed.
  • IPv6 Header (40 bytes, fixed): Version(4b), Traffic Class(8b), Flow Label(20b), Payload Length(16b), Next Header(8b), Hop Limit(8b), Source Address(128b), Destination Address(128b). No checksum (left to transport layer), no fragmentation fields, no options (extension headers instead).
  • IPv4 vs IPv6: Address: 32b vs 128b. Header: 20-60b vs 40b fixed. Checksum: Yes vs No. Fragmentation: By routers vs Source only. Configuration: Manual/DHCP vs SLAAC. NAT: Required vs Not needed. Options: Yes vs Extension headers.
  • IPv6 Address Types: Unicast (one-to-one), Multicast (one-to-many, FF00::/8), Anycast (one-to-nearest), Link-local (FE80::/10 — auto-configured, not routable). No broadcast — replaced by multicast.
  • SLAAC: Device auto-configures its address using: Router Advertisement (RA) message provides /64 network prefix. Device generates interface ID from MAC address using EUI-64 process. Combines prefix + interface ID = 128-bit global unicast address. No DHCPv6 server needed.
  • EUI-64: Take 48-bit MAC, split at middle, insert FFFE, flip 7th bit → 64-bit interface ID. E.g., MAC AA:BB:CC:DD:EE:FF → EUI-64: A8BB:CCFF:FEDD:EEFF.
Q25 · Unit II · 5/8 papers
Explain the Spanning Tree Algorithm (STA) used in bridges/switches. Why is it needed? How does it work? What are the port states in Spanning Tree Protocol (STP) and what is BPDU?
⭐ HIGH7 marksDiagram required
  • Problem: Redundant links between switches create broadcast storms (frames loop forever), MAC table instability, and duplicate frame delivery. Redundancy is desired for fault tolerance but causes loops.
  • Spanning Tree Solution: Logically disable redundant links to create a loop-free tree topology while keeping physical redundancy available. If active link fails, blocked ports can be re-enabled.
  • STP Steps: (1) Elect Root Bridge (lowest Bridge ID = Priority + MAC). (2) Each non-root bridge finds its Root Port (shortest path to root). (3) Each network segment selects Designated Port (closest to root). (4) All other ports become Blocked ports.
  • BPDU (Bridge Protocol Data Unit): Control message exchanged between bridges. Contains Root Bridge ID, Path cost to root, Sender Bridge ID, Port ID. Sent every 2 seconds (Hello time).
  • Port States: Blocking (no forwarding, receives BPDUs only) → Listening (15 sec, no forwarding) → Learning (15 sec, learns MAC addresses, no forwarding) → Forwarding (normal operation). Total convergence = 50 seconds. RSTP (Rapid STP) converges in seconds.
  • Bridge ID: 2-byte Priority (default 32768) + 6-byte MAC address. Lower value = higher priority = preferred root. Network admin can manually set lower priority to control root election.
07

Short Questions (2-mark)

Section A
ℹ️ Section A: 7 questions × 2 marks. These are definitional, one-liners, or very short explanations.
Unit II · PYQ ⭐⭐⭐
What is piggybacking?
Technique of combining an ACK with outgoing data frame to reduce overhead. The ACK "piggybacks" on the data frame traveling in the opposite direction. Reduces network traffic in full-duplex communication.
Unit I · PYQ ⭐⭐
Differentiate Hub vs Switch
Hub: Layer 1, broadcasts to all ports, shared bandwidth, no MAC table, half-duplex. Switch: Layer 2, forwards to specific port using MAC address table, dedicated bandwidth, full-duplex capable.
Unit III · PYQ ⭐⭐
What is ARP? Why is it needed?
Address Resolution Protocol maps IP addresses (Layer 3) to MAC addresses (Layer 2). Needed because Ethernet frames require physical MAC addresses. ARP broadcasts on LAN: "Who has IP X?" — device responds with its MAC.
Unit II · PYQ ⭐⭐
What is the purpose of framing in DLL?
Divides the raw bit stream into manageable data units (frames). Methods: Character count, Flag bytes with byte stuffing, Bit stuffing (inserting 0 after five 1s), Physical layer coding violations.
Unit III · PYQ ⭐⭐
What is DHCP?
Dynamic Host Configuration Protocol automatically assigns IP addresses, subnet masks, default gateways, and DNS servers. Uses DORA process: Discover → Offer → Request → Acknowledge. Port 67 (server), 68 (client).
Unit IV · PYQ ⭐⭐
What is Silly Window Syndrome?
Occurs when TCP sends/receives tiny segments repeatedly due to small window sizes, wasting bandwidth on large headers (20-byte header) for small payloads. Fixed by Nagle's Algorithm (sender side) and Clark's Solution (receiver side — advertise window only when buffer is reasonably large).
Unit I · PYQ ⭐⭐⭐
Circuit switching vs Packet switching
Circuit switching: dedicated end-to-end path established before communication, reserved bandwidth, constant delay, used in PSTN telephony. Packet switching: data split into packets, routed independently, bandwidth shared dynamically, variable delay — used in Internet.
Unit V · PYQ ⭐⭐
POP3 vs IMAP
POP3 (port 110): Downloads emails to local machine, typically deletes from server, suitable for single device. IMAP (port 143): Keeps emails on server, allows multi-device access, syncs read/unread state — modern approach.
Unit II · PYQ ⭐⭐⭐
What is CSMA/CD?
Carrier Sense Multiple Access/Collision Detection. Wired Ethernet protocol. (1) Listen before transmitting. (2) If collision detected mid-transmission — stop, send 32-bit jam signal. (3) Wait exponential backoff time. (4) Retry. Used in IEEE 802.3.
Unit III · PYQ ⭐⭐
What is ICMP and its purpose?
Internet Control Message Protocol. Error reporting and diagnostics at network layer. Examples: Type 0 Echo Reply (ping), Type 3 Destination Unreachable, Type 8 Echo Request, Type 11 TTL Exceeded (traceroute). Encapsulated in IP.
Unit IV · PYQ ⭐⭐⭐
What is QoS? Name its parameters.
Quality of Service — set of techniques to manage network resources and guarantee performance. Parameters: (1) Bandwidth — maximum data rate. (2) Latency — end-to-end delay. (3) Jitter — variation in delay. (4) Packet loss rate. (5) Reliability.
Unit V · PYQ ⭐⭐
TELNET vs SSH
TELNET (port 23): Remote terminal emulation, all data including passwords sent in plaintext — completely insecure. SSH (port 22): Secure Shell, encrypts all communication, provides authentication, tunneling, and file transfer (SCP/SFTP). SSH replaced TELNET.
Unit I · PYQ ⭐⭐⭐
What is the OSI model? Name its layers.
Open Systems Interconnection model by ISO. 7 layers (bottom to top): Physical → Data Link → Network → Transport → Session → Presentation → Application. Mnemonic: "Please Do Not Throw Sausage Pizza Away."
Unit III · PYQ ⭐⭐
What is RARP?
Reverse ARP — maps MAC address to IP address (opposite of ARP). Used by diskless workstations that know their MAC but need IP. Now obsolete, replaced by BOOTP and DHCP which provide more information.
Unit II · PYQ ⭐
What is a Bridge?
Layer 2 device that connects two LAN segments. Uses MAC address table to forward/filter frames. Unlike a hub, it doesn't flood all ports — only sends frame to segment where destination MAC resides. Reduces collision domain.
Unit I · PYQ ⭐⭐
What is multiplexing? Types?
Combining multiple signals over a single physical link. Types: FDM (Frequency Division — different frequencies for different signals), TDM (Time Division — each signal gets fixed time slot), WDM (Wavelength Division — for optical fiber), CDMA (Code Division — unique code per signal).
Unit III · PYQ ⭐⭐
What is NAT?
Network Address Translation modifies IP addresses in packet headers. Allows many private IPs (192.168.x.x) to share one public IP. Types: Static NAT (1:1), Dynamic NAT (pool), PAT/Overloading (many:1 using port numbers). Conserves IPv4 addresses.
Unit IV · PYQ ⭐⭐
What is flow control in TCP?
Mechanism to prevent sender from overwhelming receiver. TCP uses sliding window: receiver advertises its available buffer space (RWND) in ACK. Sender cannot send more data than min(CWND, RWND). Receiver can set RWND=0 to pause sender.
Unit II · PYQ ⭐⭐
What is the Hamming Code?
Error-correcting code that inserts parity bits at positions 2^k (1, 2, 4, 8...). Detects and corrects single-bit errors; detects (but cannot correct) 2-bit errors. Redundancy bits r satisfy: 2^r ≥ m+r+1 where m = data bits.
Unit I · PYQ ⭐⭐
What is a Router vs Switch?
Switch: Layer 2, uses MAC addresses, connects devices within same LAN, creates collision domain separation. Router: Layer 3, uses IP addresses, connects different networks (LANs/WANs), provides routing between networks, creates broadcast domain separation.
Unit V · PYQ ⭐⭐
What is HTTP vs HTTPS?
HTTP (port 80): HyperText Transfer Protocol, stateless, plaintext — insecure. HTTPS (port 443): HTTP over TLS/SSL, encrypts all communication with public-key cryptography. HTTPS uses digital certificates for server authentication. Now standard for all websites.
Unit III · PYQ ⭐⭐
What is subnetting?
Dividing a large IP network into smaller sub-networks (subnets) by borrowing bits from the host portion. Benefits: Reduces broadcast domain, improves security, efficient IP allocation. Subnet mask indicates which bits are network vs host. Formula: Hosts = 2^(host bits) − 2.
Unit IV · PYQ ⭐
What is the Token Bucket algorithm?
Traffic shaping algorithm. Tokens generated at constant rate r (tokens/sec). Each packet needs one token to transmit. Bucket capacity = b tokens (max burst size). Allows bursts up to b packets when tokens have accumulated. More flexible than Leaky Bucket.
Unit I · PYQ ⭐⭐
What is propagation delay vs transmission delay?
Propagation delay: Time for signal to travel from sender to receiver = Distance/Speed (~2×10⁸ m/s in copper). Transmission delay: Time to push all bits onto the wire = Frame size/Bandwidth. End-to-end delay = Tt + Tp + Queuing delay + Processing delay.
Unit V · PYQ ⭐⭐
What is SMTP? How does email delivery work?
Simple Mail Transfer Protocol (port 25) — for sending email from client to server and between mail servers. Email flow: Sender's MUA → Sender's MTA (SMTP) → Receiver's MTA (SMTP) → Receiver's MDA → Receiver's MUA (POP3/IMAP). Uses TCP for reliable delivery.
08

Key Definitions

50+ TERMS

⚡ Key Formulas Quick Reference

Shannon Capacity: C = B · log₂(1 + S/N)
Nyquist Rate: N = 2 · B · log₂(M)
Stop-and-Wait Efficiency: η = 1 / (1 + 2a), a = Tp/Tt
Pure ALOHA Throughput: S = G · e^(−2G), max at G=0.5 → 18.4%
Slotted ALOHA Throughput: S = G · e^(−G), max at G=1 → 36.8%
GBN Window Size: W ≤ 2ⁿ − 1
SR Window Size: W ≤ 2^(n−1)
RSA: C = Mᵉ mod n, M = Cᵈ mod n, e·d ≡ 1 mod φ(n)
Mesh links: n(n−1)/2 (full mesh)
Hamming distance: min distance d → detects d-1 errors, corrects ⌊(d-1)/2⌋ errors
Protocol
A set of rules and conventions governing communication between network entities. Defines format, timing, sequencing, and error control of data transmission.
Think: "Language of computers"
Bandwidth
The maximum data transfer rate of a network or channel, measured in bits per second (bps). In analog systems, it's the difference between highest and lowest frequencies of the signal.
Think: "Width of the data pipe"
Latency / Delay
The total time for data to travel from source to destination. Components: Propagation delay (distance/speed) + Transmission delay (size/bandwidth) + Queuing delay + Processing delay.
Think: "How long data takes to travel"
CRC (Cyclic Redundancy Check)
Error-detection technique using polynomial division over GF(2). Sender appends remainder (CRC bits) to data; receiver divides received frame and checks if remainder = 0 (no error).
Detects all burst errors ≤ r bits; r = degree of generator polynomial
Hamming Code
Single-error-correcting code that inserts parity bits at positions 2^k (1,2,4,8...). Formula: 2^r ≥ m+r+1 where m=data bits, r=parity bits. Can detect 2-bit errors with an extra overall parity bit.
Parity bits at positions 2^k; covers specific bit groups
Framing
Dividing continuous bit stream into discrete frames at the Data Link Layer. Methods: (1) Character count, (2) Flag bytes with byte stuffing, (3) Bit stuffing — insert 0 after five consecutive 1s, (4) Physical layer coding violations.
DLL wraps bits into addressable, manageable chunks
Piggybacking
Technique of attaching an ACK to outgoing data frame in opposite direction, rather than sending a separate ACK frame. Cuts number of frames in half for bidirectional traffic.
ACK rides piggyback on data — one frame, two purposes
CSMA/CD
Carrier Sense Multiple Access/Collision Detection. Wired Ethernet (802.3). Process: Sense medium → Transmit → If collision: send jam signal → Wait exponential backoff time → Retry. Minimum frame size = 2×Tp×Bandwidth (ensures sender is still transmitting when collision is detected).
Listen before speak, detect collision immediately, backoff
CSMA/CA
Carrier Sense Multiple Access/Collision Avoidance. Wi-Fi (802.11). Cannot detect collision while transmitting wirelessly. Uses DIFS wait + random backoff + RTS/CTS mechanism to prevent collisions proactively.
Wireless can't detect mid-transmission; must avoid beforehand
Routing
Process of selecting best path for packets from source to destination across interconnected networks. Performed at Layer 3 by routers. Static routing (manual) vs Dynamic routing (RIP, OSPF, BGP).
GPS of networking — finds best path
DVR (Distance Vector Routing)
Dynamic routing algorithm based on Bellman-Ford. Each router maintains table of distances to all destinations and shares it with directly connected neighbors periodically. Slow convergence; suffers count-to-infinity problem.
Share tables only with neighbors (not whole network) — RIP uses DVR
Link State Routing
Each router floods Link State Packets (LSPs) to entire network. All routers build identical complete topology map and independently run Dijkstra's SPF algorithm to compute shortest paths. Fast convergence.
Everyone knows everything — OSPF uses link state
ARP (Address Resolution Protocol)
Resolves IPv4 addresses to MAC addresses within same subnet. Sender broadcasts ARP Request: "Who has IP X? Tell IP Y." Device with IP X sends ARP Reply with its MAC. Result cached in ARP table (with TTL).
IP → MAC. Reverse: RARP (MAC → IP, now obsolete)
NAT (Network Address Translation)
Translates private IP addresses to public IP addresses in packet headers. PAT (Port Address Translation/Overloading) maps many private IPs to one public IP using unique port numbers. Essential for IPv4 conservation.
Many private IPs → one public IP; port numbers distinguish connections
TCP (Transmission Control Protocol)
Connection-oriented, reliable, full-duplex transport protocol. Features: error detection+recovery, flow control (sliding window), congestion control (slow start/AIMD), ordered delivery, 3-way handshake, 4-way close. Header minimum: 20 bytes.
Reliable delivery guaranteed — used for HTTP, FTP, SMTP
UDP (User Datagram Protocol)
Connectionless, best-effort transport protocol. No connection setup, no acknowledgments, no ordering, no flow/congestion control. Only 8-byte header: Source Port | Dest Port | Length | Checksum. Fast but unreliable.
Fire-and-forget — used for DNS, VoIP, streaming, gaming
Three-Way Handshake
TCP connection establishment: (1) Client → SYN(seq=x) → Server. (2) Server → SYN-ACK(seq=y, ack=x+1) → Client. (3) Client → ACK(ack=y+1) → Server. Establishes sequence numbers for both directions.
SYN → SYN-ACK → ACK = synchronized, connection open
Four-Way Handshake (TCP Close)
TCP connection termination: (1) Initiator → FIN. (2) Receiver → ACK. (3) Receiver → FIN (when ready). (4) Initiator → ACK. 2-minute TIME_WAIT state ensures all packets are flushed before socket reuse.
FIN → ACK → FIN → ACK = both directions closed
Flow Control
End-to-end mechanism preventing a fast sender from overwhelming a slow receiver. TCP receiver advertises Receive Window (RWND) in each ACK. Sender cannot have more unacknowledged data than min(CWND, RWND).
Receiver controls sender's speed via window advertisement
Congestion Control
Network-level mechanism preventing routers from being overwhelmed. TCP maintains Congestion Window (CWND). Algorithms: Slow Start (exponential growth), Congestion Avoidance (linear growth/AIMD), Fast Retransmit, Fast Recovery.
Network controls sender's speed via CWND adjustment
DNS (Domain Name System)
Globally distributed hierarchical database translating domain names to IP addresses. Hierarchy: Root (.) → TLD (.com, .in) → SLD (google.com) → Subdomain. Record types: A (IPv4), AAAA (IPv6), MX (mail), CNAME (alias), NS (nameserver).
Phone book of the internet — distributed + cached
SNMP (Simple Network Management Protocol)
Application-layer protocol (UDP 161/162) for monitoring and managing network devices. Manager sends GET/SET to Agents. Agent sends unsolicited TRAP to manager on events. MIB stores managed objects as OID tree. Versions: v1, v2c, v3 (secure).
GET=query agent, SET=configure agent, TRAP=agent alerts manager
FTP (File Transfer Protocol)
Application-layer protocol using TWO TCP connections: Control connection (port 21, commands) and Data connection (port 20 active / ephemeral passive). Active FTP: server initiates data. Passive FTP: client initiates data (firewall-friendly).
Two channels: commands on 21, data on 20 (active) or random (passive)
HTTP (HyperText Transfer Protocol)
Stateless application-layer protocol for the World Wide Web (port 80). Request-response model. Methods: GET, POST, PUT, DELETE, HEAD. HTTP/1.1 adds persistent connections. HTTP/2 adds multiplexing. HTTPS = HTTP over TLS (port 443).
Stateless client-server protocol for web pages
SMTP (Simple Mail Transfer Protocol)
Push protocol for sending email (port 25, 587). Used for: client to mail server and mail server to mail server. Commands: HELO, MAIL FROM, RCPT TO, DATA, QUIT. Contrast with POP3/IMAP which are pull (receiving) protocols.
SMTP = push outgoing mail; POP3/IMAP = pull incoming mail
Multiplexing
Sharing a single physical link among multiple signals/connections. FDM: each user gets frequency band. TDM: each user gets time slot. WDM: wavelength-division for optical fiber. CDMA: unique spreading codes. Statistical TDM: dynamic slot allocation.
Many inputs → one shared medium → demultiplex at receiver
Propagation Delay
Time for a signal bit to travel from sender to receiver through physical medium. Tp = Distance / Propagation Speed. In copper: ~2×10⁸ m/s (2/3 of c). In fiber: ~2×10⁸ m/s. Tp is fixed for a given link length.
Physical travel time of the signal — fixed for a given distance
Transmission Delay
Time to push all bits of a frame onto the link. Tt = Frame Size (bits) / Bandwidth (bps). For 1000-bit frame on 1 Mbps link: Tt = 1 ms. Different from propagation delay — Tt depends on bandwidth and frame size, not distance.
Time to pour all bits onto the wire — depends on bandwidth
Token Bucket
Traffic policing/shaping algorithm. Tokens generated at constant rate r (tokens/sec). Each packet consumes one token. Bucket capacity b = maximum burst size. Allows bursts up to b packets when tokens are saved. Output rate ≤ r over long term, but can burst to line rate short term.
Earn tokens → spend tokens to transmit → allows burst up to b
Leaky Bucket
Traffic shaping algorithm. Incoming bursty packets enter a FIFO queue (bucket). Packets leak out at constant rate r regardless of arrival pattern. If bucket overflows, packets are dropped. Completely smooths traffic — no bursts at output.
Bucket drains at constant rate — output always smooth, no bursts
Encapsulation
Process of adding headers (and sometimes trailers) at each layer as data passes down the OSI stack: Data → [TCP] Segment → [IP] Packet → [Ethernet] Frame → Bits. Decapsulation removes headers as data moves up at receiver.
Each layer wraps data in its own envelope before passing down
ICMP (Internet Control Message Protocol)
Network layer protocol for error reporting and diagnostics. Encapsulated in IP (Protocol=1). Key messages: Echo Request/Reply (ping), Destination Unreachable, Time Exceeded (traceroute), Redirect, Source Quench. Not for data transfer.
Error postcard of the internet — ping and traceroute use it
DHCP (Dynamic Host Config Protocol)
Automatically assigns network configuration to hosts. DORA process: Discover (broadcast) → Offer (server proposes IP) → Request (client accepts) → Acknowledge (server confirms). Assigns: IP, subnet mask, default gateway, DNS server, lease time.
DORA = Discover, Offer, Request, Acknowledge
Jitter
Variation in packet arrival times (delay variation). High jitter disrupts real-time applications (VoIP, video conferencing, gaming). Caused by variable queuing delays. Mitigated by playout buffers at receiver. Key QoS parameter.
Inconsistent delay — enemy of VoIP and video calls
MTU (Maximum Transmission Unit)
Maximum size of a packet/frame that can be transmitted over a network link. Ethernet MTU = 1500 bytes. If IP packet > MTU, it must be fragmented. Path MTU Discovery (PMTUD) finds smallest MTU along path to avoid fragmentation.
Largest data chunk the link can carry in one shot
Collision Domain
Network segment where all devices share the same physical medium and can cause collisions. Hubs extend collision domain. Switches and bridges break collision domains (each port = separate collision domain). Reduces with segmentation.
All devices in one collision domain compete for same medium
Broadcast Domain
Set of all devices that receive each other's broadcast frames. Switches do NOT break broadcast domains — all switch ports in same VLAN are in same broadcast domain. Routers DO break broadcast domains. Excessive broadcasts → broadcast storm.
Routers break it; switches extend it (unless VLANs used)
Store-and-Forward vs Cut-Through
Switch operation modes. Store-and-Forward: receives entire frame, checks CRC, then forwards — reliable but high latency. Cut-Through: starts forwarding after reading destination MAC (first 6 bytes) — low latency but may forward corrupt frames.
Store-forward = check then send; cut-through = send as you read
OSPF (Open Shortest Path First)
Interior Gateway Protocol using Link State algorithm (Dijkstra's SPF). Routers exchange LSAs (Link State Advertisements), build identical LSDB, compute shortest paths. Uses IP Protocol 89 (not TCP/UDP). Supports hierarchical routing with areas.
Link State protocol → complete topology map → Dijkstra's algorithm
RIP (Routing Information Protocol)
Distance Vector routing protocol. Metric = hop count (max 15 hops; 16 = infinity). Sends full routing table to neighbors every 30 seconds. Slow convergence. Suffers count-to-infinity. Simple and widely supported. Uses UDP port 520.
DVR protocol — max 15 hops, shares table every 30s
TELNET
Application layer protocol for remote terminal access over TCP port 23. All communication (including passwords) sent in plaintext — completely insecure. Uses NVT (Network Virtual Terminal) standard for character set. Replaced by SSH.
Remote terminal access — insecure (plaintext), use SSH instead
Count-to-Infinity Problem
DVR flaw: when a link fails, routers incrementally count up to infinity (by 1 each exchange) before concluding route is dead. Example: A-B-C, B-C fails. A tells B it can reach C via A (learned from B!) → loop. Fixes: split horizon, poison reverse, hold-down timers.
Routers keep counting higher thinking route exists — DVR's Achilles heel
Split Horizon
DVR loop-prevention rule: do NOT advertise a route back to the neighbor you learned it from. If router A learned "C is 3 hops via B," A should not tell B that it can reach C. Prevents routing loops and speeds up convergence.
Don't advertise a route back to where you learned it
Classful vs Classless Addressing
Classful: Fixed prefix lengths (A=/8, B=/16, C=/24). Wasteful (Class B gives 65534 hosts, often too many). Classless (CIDR): Variable-length subnet masks (VLSM), any prefix length. 192.168.10.0/26 gives 62 hosts. Classless is universal today.
Classful = rigid sizes; CIDR = flexible, efficient allocation
VPN (Virtual Private Network)
Extends private network over public network. Creates encrypted tunnel (IPSec, SSL/TLS) through internet. Types: Site-to-Site VPN (connects two offices), Remote Access VPN (employee connects to office network). Provides confidentiality, integrity, authentication.
Encrypted tunnel through public internet — acts like private link
RSA (Rivest-Shamir-Adleman)
Asymmetric public-key cryptosystem. Key generation: pick primes p,q → n=pq → φ(n)=(p-1)(q-1) → choose e coprime to φ(n) → d=e⁻¹ mod φ(n). Encrypt: C=Mᵉ mod n. Decrypt: M=Cᵈ mod n. Security: factoring large n is computationally infeasible.
Public key (e,n) encrypts; Private key (d,n) decrypts — asymmetric
09

Comparison Tables

EXAM READY
Stop-and-Wait vs Go-Back-N vs Selective Repeat
FeatureStop-and-WaitGo-Back-NSelective Repeat
Sender Window12ⁿ − 12^(n−1)
Receiver Window11 (no buffer)2^(n−1)
Efficiency1/(1+2a)W/(1+2a) if W < 1+2aW/(1+2a) if W < 1+2a
On ErrorResend 1 frameRetransmit from error frame onwardRetransmit only errored frame
Out-of-order framesN/ADiscarded by receiverBuffered by receiver
Receiver buffer1 frame1 frame onlyW frames needed
ComplexitySimplestMediumMost complex
Bandwidth useVery poor (idle while waiting)Wasteful on errorMost efficient
Pure ALOHA vs Slotted ALOHA vs CSMA/CD
FeaturePure ALOHASlotted ALOHACSMA/CD
Transmit whenAny timeAt slot boundary onlyAfter sensing medium idle
Max throughput18.4% (G=0.5)36.8% (G=1)~50% (varies with a)
Throughput formulaS=G·e^(−2G)S=G·e^(−G)η=1/(1+2a)
Vulnerable period2 × frame time1 × slot time2 × propagation time
Medium sensingNone (blind)None (blind)Yes — listen before/during
Collision handlingRandom backoff, retryRandom backoff, retryDetect + jam + exp. backoff
Used inEarly radio/satelliteRFID, some wirelessWired Ethernet (802.3)
HTTP vs FTP vs SMTP vs DNS (Application Protocols)
FeatureHTTP/HTTPSFTPSMTPDNS
Port(s)80 / 44320 (data), 21 (ctrl)25 (587 secure)53 (UDP+TCP)
TransportTCPTCPTCPUDP (TCP for zone transfer)
PurposeWeb page deliveryFile upload/downloadSend email (push)Name → IP resolution
Connections1 (persistent HTTP/1.1)2 (control + data)11
Stateless?Yes (cookies for state)No (stateful session)NoYes
DirectionPull (GET requests)Push and pullPush only (outgoing)Query/Response
SecurityTLS/SSL (HTTPS)SFTP or FTPSSTARTTLS/SMTPSDNSSEC
TCP vs UDP
FeatureTCPUDP
ConnectionConnection-oriented (3-way handshake)Connectionless
ReliabilityReliable, guaranteed deliveryUnreliable, best effort
OrderingIn-order delivery guaranteedNo ordering
SpeedSlower (overhead)Faster
Header Size20–60 bytes8 bytes
Flow ControlYes (sliding window)No
Congestion ControlYes (slow start, AIMD)No
Use casesHTTP, FTP, SMTP, SSHDNS, VoIP, streaming, gaming
Full-duplexYesYes
Error checkingYes + recoveryChecksum only, no recovery
OSI vs TCP/IP Model
FeatureOSI ModelTCP/IP Model
Layers7 layers4 layers (or 5)
DevelopmentDeveloped by ISO (theoretical)Developed by DARPA (practical)
Protocol independenceProtocol independentProtocol specific
Session/PresentationSeparate layers (5,6)Merged into Application layer
Transport protocolsNot specified by modelTCP and UDP specified
UsageReference/theoretical modelUsed in actual Internet
ReliabilityDefined at Transport layerTCP handles reliability
Go-Back-N vs Selective Repeat
FeatureGo-Back-NSelective Repeat
Sender Window2ⁿ − 12^(n−1)
Receiver Window1 (no buffering)2^(n−1) (buffering)
On ErrorRetransmit all frames from errorRetransmit only errored frame
Out-of-order framesDiscardedBuffered
EfficiencyLower (wasteful)Higher
ComplexitySimple receiverComplex receiver
BandwidthWastes bandwidth on retransmissionEfficient use of bandwidth
IPv4 vs IPv6
FeatureIPv4IPv6
Address size32 bits (4 bytes)128 bits (16 bytes)
Address notationDotted decimal (192.168.1.1)Colon-hex (2001:db8::1)
Address space~4.3 billion340 undecillion (2¹²⁸)
Header size20–60 bytes (variable)40 bytes (fixed)
NATRequired (address shortage)Not needed
FragmentationBy routers and senderOnly by sender
Security (IPSec)OptionalMandatory
BroadcastSupportedNo broadcast — uses multicast/anycast
Auto-configurationDHCP requiredSLAAC (stateless autoconfiguration)
Distance Vector vs Link State Routing
FeatureDistance Vector (DVR)Link State (LSR)
AlgorithmBellman-FordDijkstra's SPF
Information sharedRouting table with neighbors onlyLSPs flooded to entire network
Topology knowledgeLocal (only neighbor info)Complete network topology
ConvergenceSlowFast
Bandwidth usageLowHigh (flooding)
Count-to-InfinityProblem existsNo problem
ScalabilityPoor for large networksGood (hierarchical areas)
ExamplesRIP, IGRPOSPF, IS-IS
10

Flashcards

CLICK TO FLIP
0 / 40 flipped ⌨️ Arrow keys to navigate
TERM
OSI Model
Click to reveal
DEFINITION
7-layer networking framework: Physical, Data Link, Network, Transport, Session, Presentation, Application. Each layer serves the layer above and is served by the layer below.
CONCEPT
CRC Error Detection
Click to reveal
HOW IT WORKS
Divide data+zeros by generator polynomial using XOR. Remainder = CRC. Append CRC to data. Receiver divides received frame — if remainder=0, no error.
FORMULA
ALOHA Throughput
Click to reveal
ANSWER
Pure ALOHA: S = G·e^(−2G), max = 18.4% at G=0.5. Slotted ALOHA: S = G·e^(−G), max = 36.8% at G=1. Slotted is twice as efficient!
PROTOCOL
TCP Three-Way Handshake
Click to reveal
STEPS
1) Client sends SYN(x). 2) Server replies SYN-ACK(y, ack=x+1). 3) Client sends ACK(ack=y+1). Connection established! Seq numbers exchanged for both directions.
WINDOW SIZE
GBN vs SR Window
Click to reveal
ANSWER
Go-Back-N: max window = 2ⁿ−1. Selective Repeat: max window = 2^(n−1). SR window is half of GBN to avoid ambiguity at receiver. With n=3 bits: GBN=7, SR=4.
ALGORITHM
RSA Key Generation
Click to reveal
STEPS
1) Choose primes p,q. 2) n=p×q. 3) φ(n)=(p-1)(q-1). 4) Choose e coprime to φ(n). 5) Find d: e·d≡1 mod φ(n). Public key=(e,n), Private key=(d,n).
ROUTING
Count-to-Infinity
Click to reveal
PROBLEM + FIX
DVR problem: when a link fails, routers keep incrementing distance by 1 each exchange. Fix: Split Horizon (don't advertise route back to source) or Poison Reverse (advertise with ∞ cost).
PROTOCOL
DNS Resolution Types
Click to reveal
ANSWER
Recursive: DNS resolver does all work, returns final IP. Iterative: Resolver gets referrals, must query each server itself. Recursive = lazy client; Iterative = client does work.
CONGESTION
TCP Slow Start
Click to reveal
HOW IT WORKS
CWND starts at 1 MSS. Doubles every RTT until threshold (ssthresh). Then switches to additive increase (AIMD). On timeout: CWND=1, ssthresh=CWND/2.
ACCESS CONTROL
CSMA/CD vs CSMA/CA
Click to reveal
DIFFERENCE
CSMA/CD: Collision Detection — used in wired Ethernet (802.3). Detects collision while transmitting, sends jam signal, backoff. CSMA/CA: Collision Avoidance — used in Wi-Fi (802.11). Uses RTS/CTS to AVOID collision before it happens.
ADDRESSING
Subnetting /26
Click to reveal
QUICK ANSWER
Host bits = 32−26 = 6. Total addresses = 2⁶ = 64. Usable hosts = 62. Subnet mask = 255.255.255.192 (11000000 in last octet). Block size = 64.
LAYER 2
Piggybacking
Click to reveal
DEFINITION
ACK is attached to outgoing data frame traveling in opposite direction. Reduces network overhead by eliminating separate ACK frame. Used in full-duplex communication.
TRAFFIC SHAPING
Leaky vs Token Bucket
Click to reveal
KEY DIFFERENCE
Leaky Bucket: output is ALWAYS constant rate — no bursts allowed, smooths traffic. Token Bucket: allows BURSTS up to bucket size b — more flexible. Token bucket is preferred in practice.
PROTOCOL
FTP Active vs Passive
Click to reveal
DIFFERENCE
Active: Client opens control (→port 21), server initiates data connection (port 20→client). Passive: Client opens control, CLIENT also initiates data connection (server opens random port). Passive is firewall-friendly.
NETWORK LAYER
ARP vs RARP vs DHCP
Click to reveal
COMPARISON
ARP: IP→MAC (know IP, find MAC). RARP: MAC→IP (know MAC, find IP — obsolete). DHCP: Automatically assigns IP+mask+gateway+DNS to new device. DHCP replaced RARP.
EMAIL
SMTP vs POP3 vs IMAP
Click to reveal
ANSWER
SMTP (25): Send email from client to server (outgoing). POP3 (110): Download email, delete from server (one device). IMAP (143): Sync email, stays on server (multiple devices). SMTP out, POP3/IMAP in.
EFFICIENCY
Stop-and-Wait Formula
Click to reveal
FORMULA
η = 1/(1+2a) where a = Tp/Tt. If a=10: η = 1/21 ≈ 4.76% (very inefficient!). To achieve 100% efficiency with sliding window: window size W ≥ 1+2a.
MANAGEMENT
SNMP Components
Click to reveal
ANSWER
Manager (NMS): sends GET/SET queries. Agent: runs on device, responds to queries, sends TRAPs. MIB: Management Information Base — hierarchical database of managed object OIDs. Traps are unsolicited agent→manager alerts.
PHYSICAL
Shannon vs Nyquist
Click to reveal
DIFFERENCE
Nyquist: Max rate for noiseless channel = 2·B·log₂(M). Shannon: Max capacity for noisy channel = B·log₂(1+SNR). Shannon gives theoretical upper limit; Nyquist applies to clean channels.
TOPOLOGY
Mesh Network Links
Click to reveal
FORMULA
Full mesh with n nodes: links = n(n−1)/2. With 5 nodes: 10 links. Each node needs n−1 ports. Highly fault-tolerant but expensive. Used in backbone networks.
ROUTING
Dijkstra's Algorithm
Click to reveal
HOW IT WORKS
Initialize: source=0, all others=∞. Repeatedly: pick unvisited node with min cost, update neighbors if shorter path found via this node. Repeat until all visited. O(n²) complexity. Used in OSPF (Link State Routing).
PROTOCOL
DHCP DORA Process
Click to reveal
STEPS
Discover (client broadcasts) → Offer (server proposes IP, mask, gateway, DNS) → Request (client accepts offer) → Acknowledge (server confirms lease). All broadcast initially. Lease has TTL; client must renew before expiry.
NETWORK LAYER
IPv4 Header Key Fields
Click to reveal
MUST KNOW
TTL: max hops (decremented by routers, drop at 0). Protocol: upper layer (6=TCP, 17=UDP, 1=ICMP). Flags: DF (don't fragment), MF (more fragments). Fragment Offset: position in original datagram. Total size: 20 bytes minimum.
PORTS
Key Port Numbers
Click to reveal
MEMORIZE
FTP: 20/21 | SSH: 22 | TELNET: 23 | SMTP: 25 | DNS: 53 | HTTP: 80 | POP3: 110 | IMAP: 143 | HTTPS: 443 | SNMP: 161/162. Well-known ports: 0-1023. Registered: 1024-49151. Dynamic: 49152-65535.
ERROR CONTROL
CRC vs Hamming Code
Click to reveal
COMPARISON
CRC: Detection only (no correction), polynomial division, very efficient for burst errors, used in Ethernet/Wi-Fi frames. Hamming Code: Single-bit error CORRECTION, parity bits at 2^k positions, used when retransmission is costly (satellite links).
APPLICATION
DNS Record Types
Click to reveal
TYPES
A: domain → IPv4 address. AAAA: domain → IPv6 address. MX: mail server for domain. CNAME: alias (www → actual hostname). NS: authoritative nameserver. PTR: reverse lookup (IP → domain). SOA: zone authority info. TXT: text (used for SPF/DKIM).
MULTIPLEXING
FDM vs TDM
Click to reveal
COMPARISON
FDM (Frequency Division): Each channel gets unique frequency band. Used in cable TV, radio. Continuous, analog. TDM (Time Division): Each channel gets repeating time slot. Used in telephone networks, DS1/T1. Digital, efficient for bursty data. STDM adds statistical efficiency.
SECURITY
Symmetric vs Asymmetric Encryption
Click to reveal
DIFFERENCE
Symmetric: Same key for encrypt+decrypt (AES, DES). Fast, but key distribution problem. Asymmetric: Public key encrypts, private key decrypts (RSA). Solves key distribution but 100-1000× slower. In practice: Asymmetric for key exchange, Symmetric for data (TLS/HTTPS approach).
TRANSMISSION MEDIA
Twisted Pair vs Fiber Optic
Click to reveal
COMPARISON
Twisted Pair (UTP/STP): Cheap, flexible, up to 100Mbps-10Gbps, susceptible to EMI, max ~100m. Fiber Optic: Uses light, immune to EMI, very high bandwidth (Tbps), long distance (km), expensive, fragile. Types: Single-mode (long distance) vs Multi-mode (short range, cheaper).
IP ADDRESSING
IP Address Classes
Click to reveal
RANGES
Class A: 0-127 (/8) — 16M hosts. Class B: 128-191 (/16) — 65K hosts. Class C: 192-223 (/24) — 254 hosts. Class D: 224-239 (multicast). Class E: 240-255 (reserved). Private: 10.x.x.x, 172.16-31.x.x, 192.168.x.x. Loopback: 127.0.0.1.
TCP FLAGS
TCP Control Bits
Click to reveal
FLAGS
SYN: Synchronize sequence numbers (connection open). ACK: Acknowledges received data. FIN: No more data from sender (connection close). RST: Abort connection immediately. PSH: Push data to application immediately. URG: Urgent pointer field is valid. Most common in exam: SYN, ACK, FIN, RST.
ROUTING
Split Horizon
Click to reveal
RULE
DVR loop prevention rule: A router never advertises a route back to the neighbor from which it learned that route. Example: If A learned "C is 3 hops" from B, A will not tell B "C is reachable via A." Poison Reverse: actively advertise the route with metric=infinity.
NETWORK DEVICES
Hub vs Switch vs Router
Click to reveal
COMPARISON
Hub: Layer 1, broadcasts all frames to all ports, single collision domain. Switch: Layer 2, uses MAC table, per-port collision domain, same broadcast domain. Router: Layer 3, uses IP, separate collision AND broadcast domains, routes between networks. Bridge: Layer 2, like simple switch.
CONGESTION
Slow Start Algorithm
Click to reveal
HOW IT WORKS
CWND = 1 MSS initially. After each ACK: CWND doubles (exponential). When CWND = ssthresh: switch to congestion avoidance (CWND += 1 MSS per RTT). On timeout: ssthresh = CWND/2, CWND = 1, restart. On 3 dup ACKs: ssthresh = CWND/2, CWND = ssthresh (fast recovery).
ENCODING
NRZ vs Manchester Encoding
Click to reveal
DIFFERENCE
NRZ-L: 0=low, 1=high. No self-clocking. NRZ-I: Toggle on 1, keep on 0. Better for long runs of 1s. Manchester: Transition in middle of bit period (low-to-high=1, high-to-low=0). Self-clocking! Used in Ethernet 10Base-T. Differential Manchester: toggle at start on 0. Doubles bandwidth requirement.
TOPOLOGY
Star vs Bus vs Ring
Click to reveal
COMPARISON
Star: Central hub/switch, easy to add/remove, hub failure = network failure. Bus: Single coax cable, passive, single point of failure, cheap. Ring: Token-passing, each device is repeater, one failure breaks ring (unless dual ring). Tree: Hierarchical star. Mesh: Most reliable, n(n-1)/2 links for n nodes.
CHECKSUM
TCP/UDP Checksum
Click to reveal
HOW IT WORKS
Sender: sum all 16-bit words of pseudo-header + header + data. Take 1's complement = checksum. Receiver: sum everything including checksum. If result = all 1s (0xFFFF) → no error. Detects most errors but not all (can't detect all even-bit changes). Mandatory in IPv6 UDP.
IPv6
IPv6 Address Types
Click to reveal
TYPES
Unicast: One interface (Global, Link-local fe80::/10, Loopback ::1). Multicast: Group of interfaces (FF00::/8). Anycast: Nearest of a group (same address, routed to closest). NO broadcast in IPv6! Link-local addresses auto-configured from MAC (EUI-64). SLAAC = Stateless Address Autoconfiguration.
SWITCHING
Circuit vs Packet vs Message Switching
Click to reveal
COMPARISON
Circuit: Dedicated path (PSTN), constant delay, wasteful. Packet: Packets routed independently (Internet), efficient, variable delay. Message: Entire message stored-forwarded (old email), high delay. Packet switching best for bursty computer traffic. Circuit best for real-time voice.
11

Leaked Paper Analysis

2025-26
🔥 Rescheduled Paper 2025-26 — Topic Coverage
AKTU CN BCS603 · Rescheduled/Leaked session analysis based on reported question patterns

🔥 CONFIRMED HOT (2026)

  • OSI Model — full 7 layers with functions
  • Network Topologies — all types with advantages
  • CRC numerical (full step-by-step)
  • Routing Algorithms — DVR step-by-step
  • Sliding Window protocols — GBN vs SR
  • TCP/IP vs OSI comparison
  • CSMA/CD — theory + working
  • Congestion Control techniques

✅ LIKELY TO APPEAR

  • Subnetting numerical
  • DNS hierarchy + name resolution
  • FTP active vs passive modes
  • TCP header format (draw)
  • SNMP components and operations
  • IPv4 vs IPv6 comparison
  • Three-way handshake
  • Token Bucket vs Leaky Bucket

⚠️ POSSIBLE (50-50)

  • Hamming Code numerical
  • ALOHA throughput calculation
  • RSA cryptography with numerical
  • Count-to-Infinity + solutions
  • QoS parameters and mechanisms
  • SMTP email delivery process
  • ARP/RARP/DHCP comparison
  • Piggybacking concept

✗ LESS LIKELY

  • Silly Window Syndrome
  • Nagle's Algorithm
  • Shannon theorem numerical
  • VPN architecture
  • Data compression techniques
  • Network Management frameworks
  • Detailed LAN standards (802.x)
  • Bridges and Spanning Tree Protocol
📋 Year-wise Hot Topics (2021–2026)
Tracking which topics dominated each paper year

2024-25 (LATEST)

  • OSI Model + TCP/IP comparison
  • CRC numerical — given data+generator
  • Subnetting — VLSM allocation
  • TCP three-way handshake diagram
  • DNS resolution — recursive + iterative
  • Token bucket vs Leaky bucket
  • Piggybacking definition + advantages
  • SNMP protocol — components

2023-24

  • Network Topologies (all 5 types)
  • DVR routing algorithm + tables
  • GBN vs Selective Repeat efficiency
  • IPv4 vs IPv6 comparison
  • RSA algorithm with numerical
  • SMTP/POP3/IMAP differences
  • Count-to-Infinity + solutions
  • TCP header format — draw and explain

2022-23

  • OSI Model diagram + protocol examples
  • CRC error detection numerical
  • CSMA/CD working + collision backoff
  • Congestion control — slow start
  • ARP/RARP/DHCP comparison
  • FTP — active vs passive modes
  • Slotted ALOHA throughput numerical
  • QoS parameters and techniques

2021-22

  • TCP/IP model vs OSI model
  • Stop-and-Wait disadvantages
  • Hamming Code error correction
  • Dijkstra's algorithm — shortest path
  • IPv4 Header format diagram
  • DNS hierarchy explanation
  • TELNET vs SSH comparison
  • TCP vs UDP comparison table
⚠️ The above is based on PYQ pattern analysis. Actual exam questions are set by the university. Use this as a probability guide, not a guarantee.
12

80/20 Exam Strategy

PRIORITIZE
Focus on the 20% of topics that give 80% of marks. AKTU CN paper is extremely predictable.
🔥 MUST DO — 70% of paper
OSI Model — full diagram + functions + protocols per layer
Network Topologies — all 5 with pros, cons, diagrams
CRC numerical — step-by-step polynomial division
DVR Routing — Bellman-Ford table iterations
DNS — hierarchy diagram + recursive vs iterative
TCP vs UDP — 10-row comparison table
TCP Header — draw 20-byte header format
FTP — Active vs Passive mode comparison
Subnetting numerical — CIDR notation + host count
SNMP — Manager, Agent, MIB, GET/SET/TRAP
TCP Three-Way Handshake — timeline diagram
Sliding Window (GBN + SR) — window sizes + efficiency
⭐ DO IF TIME — 20% of paper
CSMA/CD + CSMA/CA — working + differences
Count-to-Infinity + Split Horizon solution
IPv4 vs IPv6 comparison
RSA algorithm — small numerical example
SMTP/POP3/IMAP — email delivery flow
Leaky Bucket vs Token Bucket
QoS parameters — bandwidth, latency, jitter, packet loss
ALOHA throughput formulas
ARP/RARP/DHCP
IPv4 Header format
✗ SKIP IF PRESSED
Silly Window Syndrome details
Nagle's Algorithm
Spanning Tree Protocol
VPN implementation details
Data compression algorithms
Detailed LAN standard internals
BGP/EGP protocols

⏱️ 3-Hour Exam Time Allocation

SECTION A (14 marks)
25 min
7 questions × 2 marks. Don't spend more than 3-4 min each.
SECTION B (21 marks)
50 min
3 questions × 7 marks. ~17 min each including diagram.
SECTION C (35 marks)
95 min
5 questions × 7 marks. ~19 min each. Draw diagrams always.
REVIEW
10 min
Check roll number, question numbers, read all questions before writing.
13

Preparation Checklist

TRACK IT
0%
0 of 35 topics completed
✓
OSI Model — all 7 layers memorized
UNIT I · CRITICAL
✓
Network Topologies — diagrams drawn
UNIT I · CRITICAL
✓
TCP/IP vs OSI comparison done
UNIT I · CRITICAL
✓
Shannon & Nyquist formulas
UNIT I · MEDIUM
✓
CRC numerical practiced (2+ times)
UNIT II · CRITICAL
✓
Hamming Code example done
UNIT II · MEDIUM
✓
Piggybacking concept clear
UNIT II · HIGH
✓
GBN vs SR window sizes memorized
UNIT II · CRITICAL
✓
ALOHA throughput formulas + numericals
UNIT II · HIGH
✓
CSMA/CD + CSMA/CA understood
UNIT II · HIGH
✓
DVR algorithm + table iteration done
UNIT III · CRITICAL
✓
Dijkstra's algorithm example done
UNIT III · HIGH
✓
Count-to-Infinity + Split Horizon clear
UNIT III · HIGH
✓
Subnetting numerical practiced (3+)
UNIT III · CRITICAL
✓
ARP / RARP / DHCP compared
UNIT III · MEDIUM
✓
IPv4 header drawn and labeled
UNIT III · MEDIUM
✓
IPv4 vs IPv6 comparison done
UNIT III · HIGH
✓
TCP vs UDP comparison (10 rows)
UNIT IV · CRITICAL
✓
TCP Header drawn from memory
UNIT IV · CRITICAL
✓
Three-way handshake diagram drawn
UNIT IV · HIGH
✓
Congestion control — slow start, AIMD
UNIT IV · HIGH
✓
Leaky vs Token Bucket explained
UNIT IV · MEDIUM
✓
QoS parameters listed
UNIT IV · MEDIUM
✓
DNS hierarchy drawn + resolution types
UNIT V · CRITICAL
✓
FTP Active vs Passive understood
UNIT V · CRITICAL
✓
SMTP / POP3 / IMAP compared
UNIT V · HIGH
✓
SNMP — Manager, Agent, MIB, TRAP
UNIT V · HIGH
✓
RSA encryption steps + numerical
UNIT V · HIGH
✓
HTTP vs HTTPS + TLS handshake
UNIT V · MEDIUM
✓
Flashcards reviewed — all 40 cards
REVISION
✓
Hamming Code numerical practiced
UNIT II · MEDIUM
✓
Split Horizon + Poison Reverse understood
UNIT III · HIGH
✓
Dijkstra's algorithm practiced (graph)
UNIT III · HIGH
✓
TCP congestion control phases memorized
UNIT IV · HIGH
✓
Read cheat sheet + all formulas revised
FINAL REVISION
14

Exam Predictions 2025-26

CONFIDENCE
🔥 Based on 8-year PYQ pattern analysis. High confidence = appeared 5+ consecutive years.
99%
OSI Model
Unit I · Sec C (7 marks)
Asked in ALL 8 papers without exception. Full diagram with all 7 layer names, functions, protocols at each layer required. Draw first, then explain. 2026 paper also had TCP/IP comparison. Budget 15-18 minutes for this one.
98%
Network Topologies
Unit I · Sec A/B/C
All 8 papers. Could be 2-mark (define one topology), 7-mark (draw all 5 with pros/cons), or comparison (Bus vs Ring vs Star). The leaked 2026 paper asked for ALL topologies with advantages and disadvantages table. Draw diagrams first.
98%
Network Topologies
Unit I · Sec A/B (2-7 marks)
Asked in all 8 papers. Either draw all 5 types (7m) or compare 2-3 specific ones (2m). Always include advantages and disadvantages. 2025-26 likely to ask Bus vs Ring vs Star.
95%
CRC Numerical
Unit II · Sec C (7 marks)
7/8 papers. Always give a data string + generator polynomial. Show full XOR polynomial division step-by-step. 2026 rescheduled paper had CRC. Practice with at least 3 different generators.
93%
DVR / Routing Algorithms
Unit III · Sec C (7 marks)
7/8 papers. Usually asked as: "Explain DVR with example" or "Apply Dijkstra's to this graph." Expect a 4-6 node graph. Show complete table iteration. Include Count-to-Infinity discussion.
92%
DNS Architecture
Unit V · Sec B/C (7 marks)
7/8 papers. Explain hierarchy (root→TLD→authoritative), recursive vs iterative resolution with diagram, DNS record types (A, MX, CNAME, NS). Caching with TTL.
90%
TCP Header Format
Unit IV · Sec B/C (7 marks)
7/8 papers. Draw 20-byte header with all fields labeled. Explain significance of each field (window size, sequence number, flags, urgent pointer). Link to TCP features in explanation.
85%
Subnetting Numerical
Unit III · Sec A/B (2-7 marks)
6/8 papers. Given IP/prefix, find: subnet mask, network ID, broadcast, host range, number of usable hosts. May also ask for VLSM (allocating different subnet sizes). Practice all /24 to /30 ranges.
85%
TCP vs UDP Comparison
Unit IV · Sec A/B
6/8 papers. Always write as a comparison table (10+ rows). Include: connection, reliability, ordering, speed, header size, flow/congestion control, use cases. Then explain when to prefer UDP.
83%
SNMP Protocol
Unit V · Sec B/C (7 marks)
6/8 papers. Explain Manager-Agent-MIB architecture. Operations: GET, GET-NEXT, SET, TRAP. Difference between SNMPv1/v2/v3 (security). UDP ports 161/162. Draw architecture diagram.
80%
Sliding Window Protocols
Unit II · Sec C (7 marks)
6/8 papers. Compare Stop-and-Wait, GBN, Selective Repeat. Include window size derivations, efficiency formulas, and numerical. Draw timeline diagrams showing frame numbers and ACKs.
15

Last-Day Cheat Sheet

PRINT THIS
🖨️ This section is optimized for printing. Use Ctrl+P → "Print Section" or print the whole page.
⚡ CRITICAL FORMULAS
Shannon: C = B·log₂(1+SNR) bps
Nyquist: N = 2·B·log₂(M) bps
S&W Efficiency: η = 1/(1+2a); a=Tp/Tt
Pure ALOHA: S=G·e^−2G; max=18.4%
Slotted ALOHA: S=G·e^−G; max=36.8%
GBN window: W ≤ 2ⁿ−1
SR window: W ≤ 2^(n−1)
Mesh links: n(n−1)/2
Subnetting: hosts = 2^(32−prefix) − 2
RSA: C=Mᵉ mod n; M=Cᵈ mod n
🔷 OSI LAYERS (Top→Bottom)
7-App: HTTP, FTP, SMTP, DNS, SNMP
6-Presentation: SSL/TLS, JPEG, MPEG, ASCII
5-Session: NetBIOS, RPC, NFS, SIP
4-Transport: TCP (port 80/443/21/22/25), UDP
3-Network: IP, ICMP, ARP, OSPF, RIP
2-Data Link: Ethernet, 802.11, PPP, HDLC
1-Physical: Bits, cables, hubs, RS-232
TCP/IP Layers: App | Transport | Internet | Network Access
🌐 IMPORTANT PORT NUMBERS
FTP: 20 (data) / 21 (control)
SSH: 22
TELNET: 23
SMTP: 25
DNS: 53 (UDP+TCP)
HTTP: 80
POP3: 110
IMAP: 143
HTTPS: 443
SNMP: 161 (query) / 162 (trap)
🔢 IP ADDRESS CLASSES
Class A: 0.0.0.0–127.255.255.255 /8
Class B: 128.0.0.0–191.255.255.255 /16
Class C: 192.0.0.0–223.255.255.255 /24
Class D: 224–239.x.x.x (Multicast)
Private A: 10.0.0.0/8
Private B: 172.16.0.0/12
Private C: 192.168.0.0/16
Loopback: 127.0.0.1
🔄 TCP FLAGS (SYN, ACK, FIN...)
SYN: Synchronize — start connection
ACK: Acknowledgment — confirms receipt
FIN: Finish — close connection gracefully
RST: Reset — abort connection
PSH: Push — deliver data immediately
URG: Urgent — urgent pointer valid
3-Way: SYN → SYN-ACK → ACK
4-Way Close: FIN → ACK → FIN → ACK
📡 KEY CONCEPTS QUICK RECALL
Piggybacking: ACK inside data frame
Silly Window: Small TCP segments — fixed by Nagle
Token Bucket: Allows bursts (≤ b tokens)
Leaky Bucket: Constant output rate always
DVR: Bellman-Ford; count-to-infinity problem
LSR: Dijkstra; floods LSPs; OSPF
ARP: IP→MAC (broadcasts on LAN)
DHCP: Auto-assigns IP on DORA process
ICMP: Error reporting (ping, traceroute)
16

Common Exam Traps

⚠️ AVOID
🚨 These are the top mistakes AKTU students make that cost easy marks. Read this the night before the exam.
✗ WrongGBN sender window = 2ⁿ
✓ RightGBN sender window = 2ⁿ − 1
Must be strictly less than 2ⁿ to prevent sequence number ambiguity. SR window = 2^(n−1) — exactly half of GBN. Example n=3: GBN=7, SR=4.
✗ WrongPure ALOHA max throughput = 36.8%
✓ RightPure ALOHA max = 18.4% — Slotted ALOHA max = 36.8%
Slotted is exactly 2× Pure. Formula: Pure S=G·e^(−2G) max at G=0.5; Slotted S=G·e^(−G) max at G=1. Never swap these in numericals.
✗ WrongARP is a Data Link Layer (Layer 2) protocol
✓ RightARP is a Network Layer (Layer 3) protocol
ARP resolves IP addresses (L3) to MAC addresses (L2). Because it uses IP addresses as input, it's classified as Layer 3 in the TCP/IP model.
✗ WrongTCP connection closes with a 3-way handshake
✓ RightTCP opens with 3-way (SYN→SYN-ACK→ACK), closes with 4-way (FIN→ACK→FIN→ACK)
TCP is full-duplex — each direction closes independently with its own FIN. TIME_WAIT state lasts 2×MSL (~2 min) after the final ACK.
✗ WrongCRC bits are prepended to data before sending
✓ RightCRC remainder is appended to the END of data
Process: data + (r appended zeros) ÷ generator = remainder (CRC). Transmit original data + CRC. Receiver divides full frame; remainder = 0 → no error.
✗ WrongShannon capacity: C = B · log₂(SNR)
✓ RightC = B · log₂(1 + SNR) — the +1 is mandatory
Also remember: SNR given in dB must be converted first → SNR = 10^(dB/10). Example: 30 dB → SNR = 1000. Missing the +1 loses full marks.
✗ WrongUsable hosts per subnet = 2^(host bits)
✓ RightUsable hosts = 2^(host bits) − 2
Subtract 2: one for the Network ID (all host bits = 0) and one for the Directed Broadcast (all host bits = 1). Both are reserved and cannot be assigned to hosts.
✗ WrongFTP port 21 is used for data transfer
✓ RightPort 21 = control/commands. Port 20 = data (Active FTP only)
Active: server connects from port 20 to client. Passive: server opens a random high port; client connects. Port 21 always carries commands like LIST, RETR, STOR.
✗ WrongSMTP is used to receive emails
✓ RightSMTP sends (pushes) email. POP3 / IMAP are for receiving
SMTP (port 25/587): outgoing mail transfer. POP3 (port 110): download + delete from server. IMAP (port 143): sync, keeps messages on server. A very common 2-mark trap.
✗ WrongRouters forward broadcast packets to all interfaces
✓ RightRouters block broadcasts — they do NOT forward them
Routers create separate broadcast domains. ARP broadcasts are always LAN-local. This boundary prevents broadcast storms from spreading across the entire internet.
✗ WrongDVR routers have a complete map of the network topology
✓ RightDVR routers only know their own distance TABLE, not the topology
DVR: share distance tables with direct neighbors only. LSR: flood Link State Packets to the entire network; each router builds a full topology and runs Dijkstra.
✗ WrongHamming code condition: 2^r = m + r
✓ RightCondition: 2^r ≥ m + r + 1 (the +1 accounts for all-zeros)
Example: m = 4 data bits. Try r = 3: 2³ = 8 ≥ 4+3+1 = 8 ✓. If 2^r < m+r+1, increase r. Hamming distance = 3 → detects 2 errors, corrects 1.
✗ WrongTCP and IPv4 headers are always exactly 20 bytes
✓ Right20 bytes is the MINIMUM — maximum is 60 bytes with options
The IHL / Header Length field (4 bits) stores length in 32-bit words. Min IHL = 5 → 5×4 = 20 bytes. Max IHL = 15 → 15×4 = 60 bytes. Applies to both TCP and IPv4.
✗ WrongLeaky Bucket and Token Bucket behave the same way
✓ RightToken Bucket allows BURSTS; Leaky Bucket always outputs at a constant rate
Token Bucket: tokens saved up to size b → burst of b packets allowed. Leaky Bucket: output rate r is always constant regardless of input burst. Use Token Bucket when bursts are acceptable.
✗ WrongDNS always uses TCP (port 53)
✓ RightDNS uses UDP port 53 for queries; TCP only for zone transfers or large responses
UDP for fast, simple name queries. TCP when response > 512 bytes, or for zone transfer (copying entire zone between authoritative servers). DNSSEC also uses TCP more.
✗ WrongStop-and-Wait efficiency η = Tt ÷ (Tt + Tp)
✓ Rightη = 1/(1 + 2a) where a = Tp/Tt — factor of 2 because ACK travels back too
Full round-trip: transmit (Tt) + propagate (Tp) + ACK returns (Tp) = Tt + 2Tp. η = Tt/(Tt+2Tp) = 1/(1+2a). The 2 is the most commonly missed part.
✗ WrongIPv4 packet fragments are reassembled at intermediate routers
✓ RightFragmentation: at routers. Reassembly: ONLY at the final destination
Intermediate routers never reassemble — too slow. IPv6 is stricter: only the source can fragment; routers drop oversized packets and send ICMPv6 "Packet Too Big".
✗ WrongCSMA/CA detects collisions just like CSMA/CD
✓ RightCSMA/CA AVOIDS collisions — it cannot detect them (hidden terminal problem)
Wireless transmitters cannot hear their own transmission clearly while sending. Solution: RTS/CTS handshake + ACK confirms delivery. Avoidance, not detection.
✗ WrongHub and Switch both operate at OSI Layer 2
✓ RightHub = Layer 1 (Physical) · Switch = Layer 2 · Router = Layer 3
Hub: dumb repeater, broadcasts bits to every port. Switch: learns MAC address table, forwards frames only to the correct port. Router: uses IP addresses to route between networks.
✗ WrongIPv6 devices use DHCP for automatic address configuration
✓ RightIPv6 uses SLAAC (Stateless Address Autoconfiguration) by default
SLAAC: device takes the /64 prefix from a Router Advertisement + generates interface ID from its MAC using EUI-64. No DHCP server needed. DHCPv6 exists for managed enterprise deployments.
CN BCS603 WAR ROOM · AKTU B.Tech Sem VI · Built for exam warriors · Good luck! 🚀
📊
📊 PROGRESS
Checklist
0%
Flashcards
0%
Unit I
0%
Unit II
0%
Unit III
0%
Unit IV
0%
Unit V
0%
⚡ Self-Test Quiz 0 / 0
Unit I · Long Q
Loading...
Section
Exit slideshow?
1 / 30
TERM
Loading...
Tap card or press Space to flip
ANSWER
← → Arrow keys to navigate · Space to flip · Esc to close
🍅
🍅 Study Timer
25:00
FOCUS
Focus session — 25 min
Sessions: