A link between routers A and B fails in a network running a link-state protocol. How does this topology change reach all other routers?
ARouter A sends an update only to its directly connected neighbors, who propagate it hop-by-hop until all routers are informed
BRouters A and B each generate a new Link-State Advertisement and flood it to every router in the network
CThe routing tables are updated centrally on a designated router, which distributes the new forwarding table to all others
DAll routers detect the failure simultaneously through periodic table exchanges and recalculate in unison
In link-state routing, the routers with the changed link generate updated LSAs and flood them to every router in the network. Flooding means each receiving router forwards the LSA out all other interfaces until the update reaches every router, which then updates its LSDB and independently reruns Dijkstra's algorithm. Option A describes distance-vector behavior (hop-by-hop propagation of routing tables). Option C describes a centralized scheme that link-state protocols explicitly avoid. Option D also describes distance-vector exchange behavior.
Question 2 Multiple Choice
Why is the count-to-infinity problem impossible in a link-state routing protocol?
ALink-state routers enforce a maximum hop count, preventing route costs from growing indefinitely
BLink-state routers use split horizon — they never advertise a route back toward the router it came from
CEvery router has a complete topology map and computes paths from first-hand link data, so there are no second-hand estimates that can loop
DLink-state protocols use triggered updates on topology changes, enabling faster convergence before loops can develop
Count-to-infinity arises in distance-vector routing because routers only know distances advertised by neighbors — if a link fails, routers may advertise incorrect paths based on stale estimates, creating loops where costs increment indefinitely. This cannot happen in link-state routing because every router has the actual network topology: each router's LSA describes only its own directly connected links, and Dijkstra runs on a complete, accurate map. There are no second-hand distance estimates to go wrong. Triggered updates (option D) help convergence speed but do not prevent count-to-infinity by themselves.
Question 3 True / False
In a link-state network, all routers run Dijkstra's algorithm independently on their own copy of the link-state database, yet their forwarding tables are globally consistent because all copies of the LSDB are identical.
TTrue
FFalse
Answer: True
This is a key insight into why link-state routing works. Because flooding ensures every router receives every LSA and builds the same topology map, each router's independent Dijkstra computation uses the same input graph. Different routers use themselves as the source node, but because they all agree on the network topology, their forwarding decisions are globally consistent — no loops or conflicting paths result. This stands in contrast to distance-vector, where consistency emerges only gradually through iterative table exchanges.
Question 4 True / False
Link-state routing is more memory-efficient than distance-vector routing because routers mainly store their own routing table rather than a complete topology database.
TTrue
FFalse
Answer: False
The opposite is true. Link-state routing requires each router to maintain a complete link-state database containing every router and every link in the network — significantly more memory than distance-vector routing, where routers store only the best distance to each destination via each neighbor. The LSDB grows with network size, which is why OSPF uses area-based hierarchy to limit the full topology each router must store. The memory and CPU cost of link-state routing is an explicit trade-off for its faster convergence and avoidance of routing loops.
Question 5 Short Answer
What is the key architectural difference between link-state and distance-vector routing that explains why link-state converges faster and avoids count-to-infinity?
Think about your answer, then reveal below.
Model answer: In distance-vector routing, each router knows only the distance to each destination via each neighbor — it has no visibility into the actual network topology. Routers share full routing tables with neighbors, so topology information propagates hop-by-hop through iterative table exchanges. This slow propagation and reliance on second-hand estimates creates the conditions for loops and count-to-infinity. In link-state routing, each router knows only its own directly connected links but immediately floods this accurate, first-hand information to every router in the network. Every router then has a complete, accurate topology map and computes shortest paths independently. Topology changes propagate in a single flooding wave rather than through iterative exchanges, enabling much faster convergence and eliminating the estimation errors that cause count-to-infinity.
The architectural contrast is: distance-vector routers know global distances but advertise only to neighbors; link-state routers know only local links but advertise globally. Flooding first-hand link information to all routers enables globally consistent computation that makes link-state reliable — each router has the same map and draws the same conclusions independently.