About This Route
Central question: How does a host decide where to send traffic when it needs to reach another host?
This route builds a host-centered network model. A host does not send traffic directly to an application idea such as “the website” or “the server code.” It uses local interfaces, IP addresses, prefixes, routes, next hops, gateways, and link-layer delivery to move traffic toward a destination.
The route is about network reachability before application protocol meaning. It explains how a machine decides whether the destination is local, which interface should be used, whether a gateway is needed, and why local link delivery is different from higher-level protocols such as HTTP or WebSocket.
Contents
- Why Reaching a Host Is Not Reaching an Application
- What a Network Means From a Host Perspective
- LANs, Local Sites, and Layer 2 Reachability
- Network Interfaces, IP Addresses, and Host Identity
- Subnets, Prefixes, and Connected Routes
- Direct Delivery, Gateways, and ARP
- Sockets as the Program Boundary Above Network Reachability
- Final Mental Model: Host, Route, Link, and Socket
Reading Path
Read the notes in order. The path starts by separating host reachability from application behavior, then builds the host’s view of networks, LANs, interfaces, addresses, prefixes, routes, gateways, ARP, and sockets.
What You Will Understand
- A host reaches another host through network-layer and link-layer decisions before any application protocol is interpreted.
- LAN, subnet, route, gateway, interface, IP address, MAC address, and socket each play different roles.
- “Same network” is not a vague physical claim. It is a routing and reachability decision made from interface configuration and routes.
Connected Reading
Builds on
- Networking as OS-Managed Input and Output. The OS chapter explains sockets as program-facing I/O endpoints, while this route explains the host and network model below that endpoint.
Extends
- How a Web Request Really Runs. This route develops the network reachability layer that the web request path depends on after request bytes leave the client.