Redundancy issues can be challenging to spot. The reason is that the network may still function, but you experience slower network behavior, usually during peak hours, and then everything goes back to normal, and so the cycle continues. You might think that a large amount of traffic during peak hours is the reason for that behavior instead of a redundancy issue.
When a network lacks a mechanism to avoid traffic loops, three things can go wrong and cause loop issues, such as:
- Broadcast storms,
- Multiple frame transmissions, and
- MAC database instability.
Broadcast Storm
Broadcast storms are typically caused by network loops, where a packet is repeatedly forwarded around in a loop, resulting in multiple copies of the packet being broadcast to all network devices. This can happen when switches are misconfigured, or a cable is plugged in incorrectly.
A broadcast storm can easily occur in a network because switches, by default, flood broadcast, multicast, and unknown unicast frames endlessly, as you can see in the example topology below.
By replicating such traffic, the network gets flooded, and the number of frames continuously rises until the switches can no longer handle them properly. This is when you experience slow network behavior, overall degraded performance, or even total network communication failure.
Multiple Frame Transmission
Multiple frame transmissions occur when multiple copies of a single unicast frame arrive at the same destination host, which can cause unrecoverable errors, as seen in the image below.
In normal network behavior, a single copy of a unicast frame should be received by the destination host. However, multiple frame transmissions are a problem without a proper loop-free mechanism in place.
MAC Database Instability
MAC database instability occurs when frames sent by a sender arrive at the destination switch through different paths and are received on different interfaces.
Consequently, the switch relearns the MAC address of the sender and rewrites that information in the MAC address table for each received copy of the same frame. This leads to storing incorrect information and utilizing most of the resources for MAC address rewrites.
As you can see in the image above, MAC database instability makes the switch update its MAC database with the wrong information, leading to inconsistencies and incorrect forwarding decisions in the future.