Click any module to expand. CN interview questions come in two forms: conceptual ("Explain TCP 3-way handshake") and numerical ("Given this IP and subnet mask, find the network address, broadcast address, and number of hosts"). This module trains both.
What networks are, the OSI and TCP/IP models, and how data flows through layers
LAN, MAN, WAN, PAN. Topologies: bus, star, ring, mesh, hybrid. Client-server vs peer-to-peer. The internet as a "network of networks." Where each topology is used in practice (star dominates modern LANs, mesh for wireless).
Physical (bits), Data Link (frames), Network (packets), Transport (segments), Session, Presentation, Application. Each layer's function, protocols, and PDU (Protocol Data Unit). Encapsulation: data → segment → packet → frame → bits. "Name the 7 OSI layers and one protocol per layer" — the opening CN interview question.
Network Interface, Internet, Transport, Application. How TCP/IP maps to OSI. Why TCP/IP is what the internet ACTUALLY uses (OSI is the reference model, TCP/IP is the implementation). The protocols that live at each layer: Ethernet, IP, TCP/UDP, HTTP/DNS/SMTP.
7 layers vs 4 layers. Theory vs practice. Session/Presentation layers merged into Application in TCP/IP. Why the distinction matters for interview answers — "which model does the internet use?" (TCP/IP) vs "which model is better for understanding?" (OSI). The most asked CN comparison question.
Circuit switching (PSTN telephone — dedicated path, guaranteed bandwidth, wasted if idle). Packet switching (internet — data split into packets, each routed independently, shared bandwidth). Message switching (store-and-forward). Why the internet uses packet switching: efficiency, resilience, cost. Virtual circuits as a hybrid.
Hub: Layer 1, broadcasts to all ports (dumb). Switch: Layer 2, forwards by MAC address (smart). Router: Layer 3, forwards by IP address (between networks). Where each is used. "Difference between hub, switch, and router" — asked in EVERY CN interview. Collision domains vs broadcast domains.
How bits travel through wires/wireless and how frames are managed on a shared medium
Analog vs digital signals. Transmission media: twisted pair (Cat5e/Cat6), coaxial, fibre optic (single-mode, multi-mode), wireless. Bandwidth, throughput, latency — the three network performance metrics. Shannon's theorem and Nyquist's theorem for channel capacity calculations.
NRZ, Manchester encoding (Ethernet uses Manchester). Multiplexing: FDM (frequency), TDM (time), WDM (wavelength — fibre optic). Why multiplexing matters: multiple signals over one medium. Statistical vs synchronous TDM.
MAC address: 48-bit hardware address (e.g., AA:BB:CC:DD:EE:FF). Ethernet frame format: preamble, dest MAC, src MAC, type/length, payload, FCS. Broadcast MAC: FF:FF:FF:FF:FF:FF. How switches use MAC address tables to forward frames. ARP: resolving IP → MAC ("What is ARP?" — a standard interview question).
Single-bit parity (detects odd errors). 2D parity (detects and corrects single-bit errors). Checksum (sum of data segments — used in IP/TCP headers). CRC (Cyclic Redundancy Check — polynomial division, used in Ethernet). Hamming distance and error correction capability. CRC numerical problems.
Stop-and-Wait: send one frame, wait for ACK. Sliding window: send multiple frames before waiting. Go-Back-N: receiver discards out-of-order frames. Selective Repeat: receiver buffers out-of-order frames. Window size and efficiency calculations. Numerical: "Given bandwidth, RTT, and frame size, calculate efficiency."
Pure ALOHA (max throughput: 18.4%) and Slotted ALOHA (36.8%). CSMA/CD: Carrier Sense Multiple Access with Collision Detection — how Ethernet works (listen, transmit, detect collision, exponential backoff). CSMA/CA: Collision Avoidance — how Wi-Fi works (can't detect collisions wirelessly, so avoid them). Why Wi-Fi is different from Ethernet at Layer 2.
IP addresses, subnetting, routing — the most numerical CN topic
Classful addressing: Class A (1.0.0.0–126.255.255.255), Class B (128.0–191.255), Class C (192.0.0–223.255.255), Class D (multicast), Class E (reserved). Default subnet masks. Private IP ranges: 10.x, 172.16–31.x, 192.168.x. Loopback: 127.0.0.1. Why private IPs exist (IPv4 exhaustion).
Subnet mask: which bits are network, which are host. CIDR notation: /24 = 255.255.255.0. Subnetting calculations: network address (AND IP with mask), broadcast address (OR with inverted mask), first/last usable host, number of hosts (2^host_bits - 2). "Given 192.168.10.50/26, find network, broadcast, and host range" — the #1 CN numerical question.
Variable Length Subnet Masking: different subnets can have different sizes. "Divide 192.168.1.0/24 into subnets for departments of 100, 50, 25, and 10 hosts." Supernetting (route aggregation): combine multiple networks into one route entry (e.g., four /24s → one /22). Reduces routing table size.
IPv4 header fields: version, IHL, TTL, protocol, source/destination IP, checksum, flags, fragment offset. TTL: prevents infinite loops (decremented at each router). Fragmentation: splitting packets when MTU is exceeded. MF flag and fragment offset. Reassembly at destination. "What is MTU?" — asked frequently.
NAT translates private IPs to public IPs. Types: Static NAT (1:1), Dynamic NAT (pool), PAT/NAT Overload (many:1 using port numbers — what home routers use). Why NAT exists: IPv4 address exhaustion. NAT breaks end-to-end connectivity — implications for peer-to-peer. "How does NAT work?" — standard interview question.
128-bit addresses (3.4 × 10^38 addresses). Address format: 2001:0db8:85a3::8a2e:0370:7334. No NAT needed (enough addresses for every device). No fragmentation at routers (done at source). No checksum (handled by upper layers). Transition mechanisms: dual stack, tunnelling, translation. Why IPv6 adoption is slow but inevitable.
ARP (Address Resolution Protocol): IP → MAC resolution. ARP request (broadcast), ARP reply (unicast). ARP cache. ICMP: error reporting (destination unreachable, time exceeded) and diagnostics (ping uses Echo Request/Reply, traceroute uses TTL). DHCP: automatic IP assignment (Discover → Offer → Request → Acknowledge). "How does DHCP assign an IP?" — frequently asked.
How routers decide the best path — algorithms and real protocols that run the internet
Static vs dynamic routing. Routing table: destination network, next hop, metric. Direct vs indirect delivery. Default route (0.0.0.0/0). Autonomous Systems (AS): groups of networks under single administrative control. Interior vs exterior routing protocols.
Distance vector algorithm: each router shares its distance table with neighbours. Bellman-Ford equation: D(x,y) = min{c(x,v) + D(v,y)}. Count-to-infinity problem. Solutions: split horizon, poison reverse, hold-down timers. RIP: max 15 hops, updates every 30 seconds. Why RIP is used in small networks and replaced by OSPF in large ones.
Link state algorithm: each router floods its link costs to ALL routers. Every router has the complete network map. Dijkstra's algorithm computes shortest path tree. OSPF: areas, Hello packets, LSA flooding, DR/BDR election. Why OSPF scales better than RIP — hierarchical design with areas. OSPF is the standard enterprise interior routing protocol.
BGP (Border Gateway Protocol): the protocol that runs the INTERNET. Inter-AS routing. Path vector: routes include the full AS path (prevents loops). eBGP (between ASes) vs iBGP (within an AS). Route selection based on policies, not just shortest path. Why BGP matters: every ISP, cloud provider, and CDN uses BGP. BGP hijacking attacks.
Reliable delivery, flow control, congestion control — how TCP makes the internet work
TCP: connection-oriented, reliable, ordered, flow/congestion control, slower. UDP: connectionless, unreliable, unordered, no flow control, faster. TCP for: HTTP, FTP, SSH, email (can't lose data). UDP for: DNS, video streaming, gaming, VoIP (speed > reliability). Port numbers: source + destination identify connections. Well-known ports: 80 (HTTP), 443 (HTTPS), 22 (SSH), 53 (DNS), 25 (SMTP).
SYN (client → server, seq=x) → SYN-ACK (server → client, seq=y, ack=x+1) → ACK (client → server, ack=y+1). Connection established after 3 packets. Why 3 packets, not 2 (prevents old duplicate SYNs from opening connections). 4-way termination: FIN → ACK → FIN → ACK. TIME_WAIT state. "Explain the TCP 3-way handshake" — THE most asked TCP question.
TCP header fields: source/dest port, sequence number, ACK number, flags (SYN, ACK, FIN, RST, PSH, URG), window size, checksum, urgent pointer. Sequence numbers track bytes, not packets. Cumulative ACKs. How TCP reassembles out-of-order segments. Why understanding the header matters for debugging with Wireshark.
Receiver advertises window size (rwnd) — "I can accept this many bytes." Sender never sends more than rwnd. Dynamic adjustment: as receiver processes data, window opens. Window size = 0 → sender stops (zero-window probing). Prevents fast sender from overwhelming slow receiver. Numerical: "Given window size and RTT, calculate throughput."
Slow Start: exponential growth (cwnd doubles every RTT) until ssthresh. Congestion Avoidance: linear growth (cwnd += 1 per RTT) after ssthresh. Fast Retransmit: 3 duplicate ACKs → retransmit immediately (don't wait for timeout). Fast Recovery: halve cwnd, avoid slow start. AIMD (Additive Increase, Multiplicative Decrease). TCP Reno vs TCP Cubic (Linux default). "Draw the TCP congestion window graph" — classic CN exam question.
Socket = IP address + port number (identifies a specific process on a specific machine). Socket pair uniquely identifies a TCP connection: (srcIP, srcPort, destIP, destPort). Ephemeral ports (49152–65535) for client connections. Socket programming concepts: bind, listen, accept, connect. Why understanding sockets matters for backend development.
The protocols users interact with — domain names, web browsing, email, and APIs
DNS hierarchy: root servers → TLD servers (.com, .org, .in) → authoritative servers (google.com). Recursive vs iterative queries. DNS record types: A (IP address), AAAA (IPv6), CNAME (alias), MX (mail), NS (nameserver), TXT (text/verification). DNS caching and TTL. "How does DNS resolution work?" — asked in every CN and web development interview.
HTTP methods: GET (retrieve), POST (create), PUT (update), PATCH (partial update), DELETE (remove), HEAD, OPTIONS. Status codes: 200 (OK), 201 (Created), 301 (Moved Permanently), 304 (Not Modified), 400 (Bad Request), 401 (Unauthorised), 403 (Forbidden), 404 (Not Found), 500 (Internal Server Error). Request/response headers. HTTP is stateless — every request is independent.
HTTP/1.1: persistent connections, pipelining (but head-of-line blocking). HTTP/2: multiplexing (multiple requests over one connection), header compression (HPACK), server push, binary framing. HTTP/3: runs over QUIC (UDP-based, not TCP) — no TCP head-of-line blocking, faster connection setup (0-RTT). Why Google invented QUIC and why HTTP/3 is the future.
Cookies: small data stored in browser, sent with every request (session ID, preferences). Sessions: server-side state tied to a cookie. Why HTTP needs cookies (it's stateless by design). REST API principles: stateless, resource-based URLs, HTTP methods as verbs. JSON as the data format. Authentication: tokens (JWT), API keys, OAuth 2.0. How modern web applications communicate.
SMTP (port 25/587): sending email (push protocol). POP3 (port 110): downloading email (deletes from server). IMAP (port 143): syncing email (keeps on server — what Gmail uses). Email flow: sender → sender's SMTP server → receiver's SMTP server → receiver's POP3/IMAP client. Why email uses multiple protocols.
The complete walkthrough: URL parsing → DNS lookup (cache → recursive → iterative) → TCP handshake → TLS handshake (if HTTPS) → HTTP GET request → server processing → HTTP response → browser rendering (DOM, CSSOM, paint). This single question tests DNS, TCP, TLS, HTTP, and browser concepts simultaneously. The ultimate CN interview question — practised step by step.
Encryption, TLS, firewalls, VPNs, and common attacks — how networks are protected
Symmetric encryption: same key encrypts and decrypts (AES, DES). Fast, used for bulk data. Asymmetric encryption: public key encrypts, private key decrypts (RSA, ECC). Slow, used for key exchange and digital signatures. Hashing: one-way (SHA-256, MD5) — used for integrity verification and passwords. "Difference between symmetric and asymmetric encryption" — standard interview question.
Digital signature: hash the message → encrypt hash with private key → receiver decrypts with public key → verify hash matches. Proves authenticity and integrity. Certificates: CA (Certificate Authority) vouches for a public key's owner. Certificate chain: root CA → intermediate CA → server certificate. Why browsers trust certain websites (certificate validation).
TLS handshake: Client Hello (supported ciphers) → Server Hello (chosen cipher + certificate) → Key Exchange (asymmetric) → Session Keys Generated (symmetric) → Encrypted Communication. HTTPS = HTTP over TLS. Why HTTPS matters: encryption + authentication + integrity. TLS 1.2 vs TLS 1.3 (1.3 is faster — 1-RTT handshake). "Explain the TLS handshake" — product company favourite.
Firewall: packet filter (Layer 3/4 — IP/port rules), stateful inspection (tracks connection state), application-level gateway (proxy — inspects content). VPN: encrypted tunnel over public internet — remote workers appear to be on the corporate network. IPSec (site-to-site VPN) vs SSL VPN (remote access). IDS/IPS: detect/prevent intrusions.
DDoS: flood server with traffic (SYN flood, amplification attacks). Man-in-the-Middle (MITM): intercept communication (prevented by TLS). DNS Spoofing: return fake IP for a domain (prevented by DNSSEC). ARP Spoofing: fake ARP replies to redirect traffic. Phishing: fake websites to steal credentials. How each attack works and how it's prevented — understanding attacks to defend against them.
How modern internet infrastructure works and practical network diagnostic tools
CDN caches content at edge servers worldwide (Cloudflare, Akamai, AWS CloudFront). User in India gets content from an Indian server, not a US server. Reduces latency, handles traffic spikes, protects against DDoS. How CDN DNS routing works (anycast). Why every major website uses a CDN. Relevant for system design interviews.
Distributing traffic across multiple servers. Layer 4 (TCP — based on IP/port) vs Layer 7 (HTTP — based on URL, headers, cookies). Algorithms: round robin, least connections, IP hash, weighted. Health checks: remove unhealthy servers from rotation. Reverse proxy (Nginx, HAProxy). Why load balancing is essential for any production web application. System design staple.
Wi-Fi standards: 802.11a/b/g/n/ac/ax (Wi-Fi 6) — frequencies (2.4 GHz vs 5 GHz vs 6 GHz), speeds, range trade-offs. CSMA/CA (why wireless can't use collision detection). Cellular: 4G LTE, 5G NR — bandwidth, latency improvements. Bluetooth, Zigbee for IoT. How mobile data reaches the internet (base station → core network → internet).
ping (ICMP Echo — is the host reachable?). traceroute (path and latency per hop). nslookup / dig (DNS resolution check). netstat / ss (active connections and listening ports). curl (test HTTP endpoints — methods, headers, status codes). Wireshark (packet capture and analysis — reading TCP handshakes, HTTP requests live). "The website is slow — diagnose the network" — using these tools systematically.
WebSockets: persistent bidirectional connection (chat apps, live data feeds). Unlike HTTP request-response, WebSocket keeps the connection open. gRPC: Google's RPC framework using HTTP/2 and Protocol Buffers — faster than REST for service-to-service communication. When to use REST vs WebSocket vs gRPC. How microservices communicate internally.
40+ problems: given IP + mask, find network, broadcast, host range, number of hosts. VLSM design exercises.
Capture live traffic, identify TCP handshakes, HTTP requests, DNS queries. See the protocols in action.
Trace every step from URL → DNS → TCP → TLS → HTTP → render. Practised until the answer is automatic.
Draw cwnd over time for slow start, congestion avoidance, fast retransmit/recovery. Identify ssthresh changes.
Given a "broken" network: use ping, traceroute, dig, netstat, curl to diagnose and identify the problem.
Make API calls using curl: GET, POST, PUT, DELETE. Read headers, status codes, response bodies. Practical HTTP understanding.
Trainers explain protocols with real packet traces, not just diagrams. "Here's what a TCP handshake ACTUALLY looks like in Wireshark." Every concept anchored to observable reality.
Subnetting, sliding window efficiency, TCP throughput, channel capacity — CN has as many numericals as OS. Worked examples + practice sets for numerical topic.
Capture and analyse real network traffic. Identify protocols, trace connections, diagnose problems. Seeing TCP/HTTP/DNS in action makes abstract concepts concrete.
MCQs on concepts (OSI layers, TCP vs UDP, subnetting) + numerical problems (subnet calculation, efficiency). Auto-evaluated with topic-wise analytics.
TCS NQT, Infosys SP, Wipro, Cognizant — every service company includes CN MCQs. OSI layers, TCP vs UDP, subnetting, HTTP status codes, DNS. Students who skip CN lose marks in sections that DBMS/OS prep can't recover. Master these fundamentals to secure your campus placement offer.
This single question tests DNS, TCP, TLS, HTTP, and browser rendering — all at once. It's asked at Google, Amazon, Microsoft, and virtually every product company. A complete, confident answer demonstrates integrated CN understanding. This module practises it until the answer is automatic.
"Design a URL shortener" "Design a chat system" — every system design question requires understanding of: DNS, load balancers, CDNs, HTTP/WebSockets, TCP/UDP, caching, and network latency. CN is the foundation of system design answers — without it, you can't discuss architecture trade-offs.
Backend developers debug HTTP responses. DevOps engineers configure load balancers. Cloud engineers set up VPNs and firewall rules. Full-stack developers make REST API calls. CN isn't just an interview topic — it's daily work knowledge for every technical role that involves the internet (which is all of them).
E-305, Industrial Area, Sahibzada Ajit Singh Nagar, Punjab 160055, IN
