New: draw a network as text and trace a packet through it hop by hop.Open the routing simulator

Topology & Routing Simulator

betaFrom vendor documentation

Describe a network as text, then trace a packet through it hop by hop: connected, static and single-area OSPF routes, longest-prefix forwarding, equal-cost multipath, and the exact hop where it dies when it does not arrive.

Topology definition

6 devices · 7 links

What this simulates, and what it does not

Connected and static routes, and single-area OSPF reduced to the state it settles into: a shortest-path tree with equal-cost multipath. Forwarding is a real longest-prefix match against the resulting table, hop by hop, including the layer-2 path across switches.

It does not simulate convergence — no hello timers, no DR/BDR election, no LSA flooding, no areas, no BGP, no spanning tree, no ARP ageing. It answers “given that routing has settled, where does the packet go?” and it cannot answer “how long does that take?”

Diagram

Drag a device to move it — the coordinates are written back into the text

10.0.12.0/3010.0.13.0/3010.0.14.0/3010.0.24.0/3010.0.34.0/30r1routerr2routerr3routerr4routerpc1hostsw1switch

An address, or a device name

r1

Routing table and interfaces

Result
Delivered
Hops
3
Source address
10.0.12.1

Reachable

Delivered to 4.4.4.4 on r4:Lo0.
  1. 1r1forwardospfAD 110metric 214.4.4.4/32 [ospf metric 21] → Gi0/0 via 10.0.12.2 (2 equal-cost paths)
  2. 2r2forwardospfAD 110metric 114.4.4.4/32 [ospf metric 11] → Gi0/1 via 10.0.24.1
  3. 3r4deliverDelivered to 4.4.4.4 on r4:Lo0.

Frequently asked

Does this replace a lab like GNS3 or EVE-NG?
No, and it does not try to. It simulates the forwarding plane: given a topology and the routes those protocols settle on, where does a packet actually go? It runs no protocol, so it cannot tell you how a network converges, only where traffic ends up once it has.
What exactly is simulated?
Connected routes derived from interface addresses, static routes, and single-area OSPF reduced to its steady state — a Dijkstra shortest-path tree over the interfaces that could form an adjacency, with equal-cost multipath preserved. Forwarding is a real longest-prefix match, then administrative distance, then metric, hop by hop, including the layer-2 path across switches.
Why does a hop say ARP failed?
Because the route was right and nobody owned the address. When the best match is a connected route the packet is on the right wire, and the simulator looks for the device configured with that exact address. Finding none, it reports what a real network would do: nothing answers.
Why did OSPF not install a route over a link I configured?
Most often the two interfaces are not in the same subnet, so the adjacency would never form. The routing table reports that as a warning rather than quietly omitting the route, because a silent omission looks exactly like a correct answer.
What is equal-cost multipath doing here?
When two paths to the same prefix cost the same, both next hops are installed and the tool shows both. A simulator that kept only one path could not answer the question people actually ask about a topology — why traffic is using the link it is using.