Network routing

A router uses a destination address and forwarding information to choose an outgoing direction. It usually selects a next hop rather than requiring the application to describe an entire journey for every packet. Routing supplies the information behind those choices, while forwarding applies established rules to individual arrivals at a finer timescale.

References: [1]

Routes may come from static configuration or protocols exchanging connectivity and reachability. OSPF distributes link-state information and computes paths from a topology description. Neighbor relationships, updates, and aging allow routers to revise earlier decisions when connectivity changes, instead of depending permanently on the table present at initial installation.

References: [2]

Path comparisons use defined metrics, and administrative domains can impose policy. A shortest path is shortest according to its chosen cost, not necessarily geographical distance or measured delay. Availability, resource use, and management requirements influence the decision, so apparently nearby destinations can legitimately leave a network through different gateways.

References: [1]

After a failure, routers can learn new information at different times, causing temporary loops or unreachable destinations. Packet lifetime limits bound looping, while routing protocols work toward consistent paths. Diagnosing a failed delivery requires routing state and link conditions together, distinguishing persistent configuration problems from transient behavior during convergence.

References: [2]