Internet Routing: Distance-Vector vs Link-State (2D)

Algorithm

Distance-Vector (RIP-like) Link-State (OSPF-like)

Source & destination

Source router

Destination router

Topology

Canvas edit mode

Animation speed 5

Live readout

Nodes / links

0 / 0

Rounds to converge

—

Best-path cost

—

Hop count

—

Ready.

How it works

Distance-Vector (Bellman-Ford, like RIP): each router only tells its direct neighbours its own distance table, every round, synchronously. A router updates its table whenever neighbourDist + linkWeight beats what it already has. When a link that was in use fails, a stale advertisement can bounce back and forth, slowly inflating the distance instead of dropping it — the classic count-to-infinity problem, capped here at 16 the way RIP caps it.

Link-State (Dijkstra, like OSPF): every router floods the full topology to everyone first, so all routers end up with an identical map, then each runs Dijkstra's shortest-path algorithm independently in one pass — no multi-round back-and-forth, and no count-to-infinity.

Use Cut / restore link mode and click any edge to break it, then Step or Auto-run to watch each algorithm reconverge — Distance-Vector visibly struggles more on sparse topologies (Ring, Star) than Link-State does.

Drag a router to move it · click a link to cut/restore it