Routing aims to find paths that optimize metrics such as shortest hop count, latency, loss, bandwidth, policy compliance, and resilience. In practice, routing is constrained by incomplete information, dynamics, administrative policy, and scalability limits.

Intra-domain routing such as link-state and distance-vector protocols balances convergence speed, overhead, and loop freedom. Link-state protocols disseminate topology broadly, while distance-vector protocols rely on local neighbor information and iterative updates.

Inter-domain routing is shaped by policy, trust, and economic relationships rather than purely technical shortest-path criteria. Path selection can be constrained by business agreements, valley-free rules, and policy filtering.

Important open problems include fast convergence under churn, scalable policy inference, secure route validation, and routing under partial information. Multipath and traffic-engineered routing also raise questions of stability and fairness.