Computer Networks & Architecture

1. The layered architecture of networks

Layering is the single design decision that made the internet grow: each layer offers a service to the layer above and uses the layers below without knowing how they work. This lesson covers the OSI and TCP/IP models properly — not as a diagram to memorise but as an engineering tool — and shows what the separation buys you: interchangeable technologies, independent security boundaries and interfaces that can be reasoned about one at a time.

2. Routing and switching internals

A router does not know the way; it knows one next hop at a time, chosen by the longest matching prefix in a table. A switch does not route at all; it learns which MAC addresses live on which ports and floods anything it does not know. This lesson covers forwarding decisions, the routing table, the protocols that fill it (OSPF, BGP), switching and VLANs, NAT in practice, and the MTU and fragmentation problems that produce the strangest bugs in production.

1. Subnetting, segmentation and redundancy

A network design is mostly two decisions repeated: how you slice the address space, and what happens when a piece fails. This lesson makes subnetting arithmetic routine — masks, usable ranges, host counts, VLSM planning — then covers segmentation as a security control, redundancy and the trade-off between availability and complexity, and capacity planning based on measured traffic rather than vendors' claims.

2. Network performance, monitoring and failure

The last lesson of the course is about evidence: how to measure a network so that arguments end, how throughput, latency, jitter and loss interact with TCP's own behaviour, how to monitor a network for the failures that matter, and how modern topologies — spine-and-leaf, SDN, CDNs and cloud networks — change the way performance is designed rather than repaired.