What Does It Mean for a Routing System to Be "Redundant," and Is Redundancy Good or Bad?

The answer

When a routing system is described as redundant, it means the network is built with more than one path between endpoints. Instead of a single line connecting point A to point B, there are backup links, backup routers, and alternate routes. If any one component fails, the routing protocol detects the failure and shifts traffic onto a surviving path, usually within seconds and without a human touching anything.

Redundancy is a good thing. Its whole purpose is to eliminate single points of failure. A network with only one path is fragile: cut that one link, or lose that one router, and everything downstream goes dark. Add a second, independent path and the same failure becomes a non-event because traffic simply detours around the broken part. This property is called fault tolerance, and it is why banks, hospitals, airlines, and internet backbones all pay for redundant designs.

How rerouting actually works

Redundancy only helps if the network can notice a failure and react to it. That job belongs to dynamic routing protocols such as OSPF, EIGRP, and BGP. Each router constantly exchanges reachability information with its neighbors and keeps a routing table of the best-known paths. When a link goes down, neighboring routers stop receiving the expected "hello" or keepalive messages, mark that route as unavailable, and recalculate. The next-best path from the table takes over. This convergence happens automatically, which is the payoff of building redundancy in the first place.

Protocols like HSRP or VRRP add redundancy at the gateway level: two routers share a single virtual IP, and if the active one dies, the standby instantly assumes the role so devices never notice.

Why redundancy has a cost, not a downside

Redundancy is not free, and that is the honest caveat behind the "is it good or bad" question. More paths mean more routers, more cabling or circuits, and more money spent on equipment you hope never has to carry live traffic. It also adds configuration complexity: engineers must prevent routing loops, tune failover timers, and test that backups genuinely work. A poorly configured redundant network can even create new problems, like broadcast storms, if safeguards such as Spanning Tree Protocol are missing.

But these are trade-offs, not reasons to call redundancy bad. For any system where downtime is costly or dangerous, the added expense and complexity are well worth the reliability gained.

Redundancy versus load balancing

Students often confuse redundancy with load balancing. They overlap but are not the same. Redundancy is about availability -- keeping a backup ready so the network survives a failure. Load balancing is about performance -- spreading traffic across multiple active paths so no single link is overwhelmed. A well-designed network frequently does both: multiple paths carry traffic during normal operation (load balancing) and also serve as each other's backups (redundancy).