Topic 4c – Internet Protocols and Client-Server Interaction

Packets, Layers, and the Webpage Journey

1. A packet is a small, fixed-size chunk of data with a header containing the source address, destination address, and sequence information. The Internet breaks messages into packets because: (1) it allows multiple conversations to share the same network links simultaneously (packet switching); (2) if one packet is lost or corrupted, only that packet needs to be retransmitted, not the whole message; (3) packets can take different routes around congestion or failures, making the network more resilient.
2. Packets from the same message can take different routes because routers make independent forwarding decisions based on current network conditions — if one path is congested, a router may send subsequent packets a different way. The destination knows how to reassemble them correctly because each packet carries a sequence number in its header. The receiving device collects all packets, reorders them by sequence number, and reassembles the original message. TCP also handles retransmission of any missing packets.
3. The four layers: (1) Application layer — the protocol used by the specific application (HTTP, SMTP, DNS); present only at origin and destination. (2) Transport layer — TCP or UDP; manages end-to-end delivery, ports, reliability; present only at origin and destination. (3) Internet/Network layer — IP addressing and routing; present at every device including routers. (4) Link layer — local delivery on a single physical network segment; present at every device. Routers operate primarily at the Internet layer (and link layer) but not the transport or application layers.
4. Sample trace (five steps minimum):
   • DNS lookup (Application/Transport/Internet layers) — browser sends a DNS query to resolve the domain name to an IP address.
   • TCP handshake (Transport layer) — browser establishes a TCP connection with the web server at that IP address.
   • HTTP request (Application layer) — browser sends an HTTP GET request for the specific page. \
   • Packet routing (Internet/Link layers) — the request is broken into packets, each routed across the Internet through multiple routers.
   • HTTP response (Application layer) — the web server sends back the HTML and associated resources in packets.
   • Reassembly and rendering (Transport/Application layers) — the browser reassembles the packets and renders the HTML into a visual page.
5. In circuit switching (used by traditional telephone networks), a dedicated physical path is reserved between caller and receiver for the entire duration of the call — even during silences, that capacity is occupied. Packet switching sends data in independent packets that share network links with everyone else's traffic. The advantage: packet switching makes much more efficient use of network capacity, since links are only used when data is actually being sent. This is why the Internet can serve billions of users simultaneously on shared infrastructure.

Ports

6. An IP address identifies a device on the network, but a single device may be running dozens of different network applications simultaneously (a browser, an email client, a video call, a game). Port numbers identify which application on that device should receive a given packet. Without ports, the operating system would have no way to route an incoming packet to the correct process.
7. The incoming packet's header contains a destination port number. The transport layer reads that number and delivers the packet to whichever application is currently listening on that port. HTTP traffic with destination port 80 goes to the web server process; HTTPS traffic with destination port 443 also goes to the web server (its HTTPS listener); SMTP traffic with destination port 25 goes to the mail server process.
8. The :8080 indicates a non-standard port number — the web server for this particular resource is listening on port 8080 instead of the default port 80. This is common when a second or test web server is running on the same machine alongside the main one (which uses port 80), or when a standard web server on a school network is blocked by the firewall. Port 8080 is a widely used alternative HTTP port.

TCP vs. UDP

9. Connection-oriented (TCP) means TCP establishes a connection before any data is sent (via the three-way handshake), guarantees delivery and ordering, and formally closes the connection when done. Connectionless (UDP) means UDP sends packets with no prior setup, no acknowledgment, and no guarantee that they arrive or arrive in order. Practically: TCP applications receive complete, correctly ordered data but at the cost of latency; UDP applications receive data as fast as possible but may experience gaps or glitches.
10. "Unreliable" does not mean UDP is broken — it means UDP does not include the overhead of guaranteed delivery. For applications where speed matters more than perfection — live video calls, online games, streaming audio — a slightly dropped packet causes only a brief glitch, while TCP's retransmission overhead would cause a noticeable delay or stutter. UDP is often the better choice precisely because it does not try to fix lost packets: by the time a retransmission arrives, the data would be stale anyway.
11. • Loading a webpage: TCP — all HTML, CSS, and images must arrive correctly and completely for the page to display properly.
    • Live video call in Google Meet: UDP — real-time audio/video can tolerate occasional lost packets (brief glitches) but cannot tolerate the delays caused by TCP retransmission.
    • Downloading a software update: TCP — every byte of the installer must arrive correctly; a corrupted file would be unusable.
    • A DNS lookup: UDP — DNS queries are tiny and need fast responses; if no reply arrives, the client simply re-sends the query. The overhead of TCP connection setup would be disproportionate.
    • Sending an email: TCP — an email must arrive complete and intact; SMTP uses TCP to ensure reliable delivery between mail servers.
12. The speed advantage is worth the risk for the right applications. When you are on a live video call, a packet that arrives 200 milliseconds late is worse than useless — TCP would delay all subsequent audio waiting for it, causing a noticeable stutter. UDP simply moves on; the receiver fills the gap with a brief silence or glitch and continues. For real-time applications, the latency cost of reliability is higher than the quality cost of occasional loss. UDP is not a flaw; it is a deliberate design choice for scenarios where speed beats completeness.