TechKnowSurge
Cisco CCNA 3.4 Cisco CCNA 3.2
VideoNetworkFree

Load Balancing

Network redundancy and load balancing allow routers to distribute traffic across multiple equal-cost paths, maximizing bandwidth and providing automatic failover when a route goes down. OSPF supports up to four equal-cost paths by default and up to sixteen when configured manually, while EIGRP extends this further with unequal-cost load balancing across links of different speeds.

Complete this video to capture a CTF flag worth 1 point.

About this video

Network redundancy is a foundational design principle that ensures connectivity is maintained even when individual links or devices fail. Rather than relying on a single path between two points, well-designed networks establish multiple routes so that traffic can automatically fail over to an alternative if the primary link goes down. When multiple paths exist, those links can also be leveraged simultaneously through load balancing, distributing packets across routes to increase aggregate throughput to a destination. OSPF performs equal-cost load balancing automatically when two conditions are met: the administrative distance and the metric, or cost, must be identical across all candidate paths. When those values match, OSPF will distribute traffic across up to four routes by default. That limit can be adjusted using the maximum-paths command within the OSPF process configuration, allowing anywhere from a single forced path up to a maximum of sixteen simultaneous routes. EIGRP offers an additional capability that OSPF does not: unequal-cost load balancing. Where OSPF requires all paths to share the same metric, EIGRP can distribute traffic across links with different speeds or costs, sending proportionally more traffic over higher-bandwidth connections while still utilizing lower-bandwidth links. This makes EIGRP particularly flexible in environments where redundant paths are not symmetrical, allowing network engineers to extract usable capacity from every available link rather than leaving slower paths idle.

What you'll learn

What's covered

Network Redundancy & Load Balancing

Aligned to

Cisco CCNA
3.4 Configure and verify single area OSPFv2
3.2 Determine how a router makes a forwarding decision by default

Key terms

Redundancy
The duplication of critical components or systems to increase reliability and availability.
Failover
The automatic switching to a redundant system or component when the primary one fails.
Routing Protocol
A protocol used by routers to communicate and share routing information to determine the best path for data.
Open Shortest Path First
OSPF
A link-state interior gateway routing protocol that maintains a complete map of the network topology and uses Dijkstra's Shortest Path First algorithm to calculate optimal, loop-free routes. OSPF divides networks into hierarchical areas to improve scalability and supports VLSM and rapid convergence.
Equal-Cost Load Balancing
A routing feature that distributes traffic across multiple paths that share the same administrative distance and metric.
Active Directory
AD
Microsoft's directory service used to manage users, computers, and resources in a network.
OSPF Metric
The cost value OSPF assigns to a route based on bandwidth; routes with equal cost are eligible for load balancing.
Maximum Paths
An OSPF configuration parameter set under the router OSPF process that controls the number of equal-cost routes used for load balancing, with a default of 4 and a maximum of 16.
Enhanced Interior Gateway Routing Protocol
EIGRP
A Cisco advanced distance-vector routing protocol that uses the DUAL algorithm to achieve rapid, loop-free convergence. EIGRP supports VLSM, unequal-cost load balancing, and sends partial updates only when topology changes occur, conserving bandwidth compared to traditional distance-vector protocols.
Unequal-Cost Load Balancing
A routing feature supported by EIGRP that distributes traffic across multiple paths with different metrics, sending more traffic over higher-bandwidth paths.

Transcript

One thing we want within our networks is to have some sort of redundancy. We want redundancy with our equipment, we want redundancy within our equipment, and we want redundancy between our equipment. So with these links, we want multiple routes to go to the same location, and these are for when one route dies and we want it to fail over to the other route. But there are some capabilities of doing load balancing. We have multiple routes going to the same location, so why do we not use multiple routes to get our data to that location, giving us more bandwidth?

Load Balancing With OSPF

Here is an example of a network that has two different routes that it can take to get to this network over here. This router right here can do load balancing to actually increase performance. What it will do is send some of the packets in this direction and other packets in this direction, giving it twice as much bandwidth to its end destination.

Now with OSPF, two things need to be the same. The administrative distance, which is pretty simple because administrative distance would be 110 on both of those for OSPF, and then also the metric, so the cost has to be the same whether it is going this direction or whether it is going this direction. As long as these two are the same, the metric and the AD are the same, your router will actually automatically load balance between these two different routes.

In fact it will actually load balance up to four different routes by default, provided that the AD and the metric are the same. By default OSPF, as long as the administrative distance and the metric are the same, can load balance up to four, and that is just by default. It can actually go up to 16 different routes that it can load balance between all of these different routes. Other routing protocols have similar limits to them and similar defaults to them.

Changing The Maximum Number Of Paths

So how can we change this? How can we utilize more than what the default is? If we go into the router ospf process ID and then type in maximum-paths and then the number of paths that we want to take, so let us say we want to do all 16 and we type in 16 there, then we would actually change this default to something much greater. Or less, possibly. We may want to do less, maybe we want to choose just one route, and we would just specify that the maximum number is one, so that way it will not load balance. So those are some ways that we can actually get more out of our different paths to the end destination.

Unequal Cost Load Balancing

Just a side note that EIGRP can actually do unequal cost load balancing. What that means is if these are 10 gig connections over here and these are one gig connections over here, it could utilize both of these connections, and it would utilize this one much more than it would utilize this connection. So that is one of the powers behind EIGRP: it can do unequal cost load balancing.

About TechKnowSurge

TechKnowSurge builds IT and cybersecurity professionals through hands-on, concept-first training built around real understanding — not memorization. Free interactive tools, structured programs, and 25+ years of real-world experience, all in one place.

Explore free tools and programs →