Computer Networking

A Top-Down Approach — an interactive study site

Built from Computer Networking: A Top-Down Approach, 8th edition (2021), by James F. Kurose (University of Massachusetts, Amherst) and Keith W. Ross (NYU), published by Pearson.

This site is a personal, non-commercial study aid derived from that book. It is not the book, and it is not affiliated with the authors or the publisher. All explanations, diagrams and exercises here are rewritten, not reproduced.

What is here

121 of 121 pages built
8
chapters
121
section pages
748
quiz questions, each option explained
690
book pages covered

The book as a protocol stack

The book works top-down: it starts at the application layer, where you already have intuition, and digs one layer down at a time.

Before the stack — the whole picture

1. Computer Networks and the Internet
Application layer
Transport layer
Physical layer
covered inside chapters 1 and 6 (physical media, encoding)

The same book as a learning path

The stack above says where a chapter sits. This says what it needs first — and the two orders are not the same. Chapter 6 is at the bottom of the stack but draws on four earlier chapters; chapter 5 sits in the middle and needs essentially only chapter 4. Every line below is taken from that chapter's own "depends on" notes, written while it was being built.

  1. 1. Computer Networks and the Internet assumes no networking knowledge — start here · 6 later chapters come back to it
  2. 2. Application Layer needs 1 earlier chapter · 4 later chapters come back to it
    • ← ch 1 §1.5, §1.1, §1.4, §1.3.3 The five layers and encapsulation; protocols and the socket interface; the delay components §2.2.5 reuses directly; the network of networks §2.6 builds CDNs on.
  3. 3. Transport Layer needs 2 earlier chapters · 5 later chapters come back to it
    • ← ch 2 §2.1.3–2.1.4, §2.7, §2.2.2, §2.2.5–2.2.6, §2.4, §2.6.3 Chapter 2 asked what TCP and UDP promise an application; this chapter answers it. §3.2 re-reads chapter 2’s own socket code from the transport layer’s side.
    • ← ch 1 §1.4.2, §1.4.4, §1.5, §1.6 Queuing delay and traffic intensity — §3.6’s three scenarios are that one graph, three times. The bottleneck link becomes §3.7.3’s fairness argument.
  4. 4. The Network Layer: Data Plane needs 3 earlier chapters · 4 later chapters come back to it
    • ← ch 3 §3.1, §3.5.2, §3.6–3.7 The network layer’s service model is a constraint on transport, and this chapter is the other side of it. ACK clocking is what produces bufferbloat in §4.2.4.
    • ← ch 1 §1.3, §1.4.1–1.4.2, §1.5.2 Packet switching and store-and-forward — §4.2 is the inside of that picture, and §4.2.4 is where the queueing delay actually happens.
    • ← ch 2 §2.6 Content distribution networks, the ancestor of §4.5’s performance-enhancing middleboxes.
  5. 5. The Network Layer: Control Plane needs 4 earlier chapters · 2 later chapters come back to it
    • ← ch 4 §4.1, §4.2.1, §4.3.2, §4.4, §4.5 This chapter is the half chapter 4 deferred: Figures 5.1 and 5.2 are Figures 4.2 and 4.3 redrawn. Everything it computes ends up in a forwarding table.
    • ← ch 1 §1.3.3 The access / regional / tier-1 ISP hierarchy, which §5.4’s BGP policy scenarios are stated in terms of.
    • ← ch 2 §2.6 CDNs — §5.4.4’s IP-anycast is how one steers users to a nearby replica.
    • ← ch 3 §3.7.2 ECN and congestion; §5.2.1’s route oscillation is congestion-sensitive routing misbehaving.
  6. 6. The Link Layer and LANs needs 4 earlier chapters · 1 later chapter come back to it
    • ← ch 4 §4.2, §4.3.1, §4.3.2 The IPv4 datagram and the MTU — Ethernet’s 1,500-byte payload is why the MTU is 1,500. §6.4.3 contrasts a switch with a router point by point.
    • ← ch 3 §3.3, §3.4 The Internet checksum, which §6.2.2 contrasts directly with CRC; and ARQ, which §6.2 contrasts with forward error correction.
    • ← ch 1 §1.5, §1.3.1, §1.4, §1.2.1 The layer below everything so far. d_prop and d_trans are reused unchanged in the CSMA/CD efficiency formula.
    • ← ch 5 §5.5 SDN — §6.6.2’s data-centre trends are SDN in production.
  7. 7. Wireless and Mobile Networks needs 3 earlier chapters · 1 later chapter come back to it
    • ← ch 6 §6.3, §6.3.1 The whole multiple-access framework. §7.3.2 is best read as "why chapter 6’s answer does not work here", and chapter 6 deliberately deferred CDMA to §7.2.
    • ← ch 4 §4.3, §4.3.3 Addressing and DHCP, for association and care-of addresses; and tunnelling, which is exactly the mechanism indirect routing uses.
    • ← ch 3 §3.5, §3.7 TCP’s assumption that loss means congestion — which §7.7 attacks.
  8. 8. Security in Computer Networks needs 6 earlier chapters
    • ← ch 7 §7.3–7.4 §8.8 secures exactly the 802.11 and 4G/5G attachment procedures chapter 7 built.
    • ← ch 1 §1.6 The attack taxonomy this chapter finally answers: malware, denial of service, sniffing, masquerading, message modification.
    • ← ch 3 §3.5 Checksums — §8.3 opens by explaining why a checksum is not a cryptographic hash.
    • ← ch 4 §4.3.1 The IPv4 datagram, which §8.7’s tunnel mode wraps whole.
    • ← ch 2 §2.4 DNS, named as infrastructure that itself needs secure communication.
    • ← ch 5 chapter 5 Routing and network management, likewise named as infrastructure needing protection.

Reference pages