Switching loop
(also bridge loop, looped network)
Switching loop definition
A switching loop is a network condition where the interconnections between switches and bridges create a looped path for data packets. Switching loops occur when there is more than a single layer 2 (the data link layer) path between two endpoints.
Switching loops can lead to high CPU usage on switches, excessive network traffic, network outages, and slow response times. Because layer 2 headers of data frames do not have time to live (TTL) values, packets can circulate forever once a switching loop develops.
How a switching loop works
When a loop is formed due to an unintentional interconnection between bridges or switches, data packets sent from one segment start to loop endlessly without reaching their intended destination. As they circulate, they are repeatedly flooded to all segments connected to the looped path, consuming network resources.
Eventually, the accumulation of redundant packets leads to a broadcast storm. The storm can overload layer 2 connecting nodes and cause packet collision, resulting in the network going down.
Stopping switching loops
- Implement proper network architecture with a hierarchical structure, avoiding redundant physical connections between switches.
- Use the Spanning Tree Protocol, which is designed to prevent switching loops. The Spanning Tree Protocol creates a loop-free topology within the network and blocks redundant paths.
- Use the Rapid Spanning Tree Protocol, which is an improved version of the Spanning Tree Protocol with faster convergence and recovery times. It detects network changes more quickly and allows for faster reconfiguration of the network topology.
- Use Loop Guard to detect and prevent switching loops caused by the absence of bridge protocol data units (BPDU) on forwarding ports. Loop Guard monitors the state of the designated forwarding port and places it into a loop-inconsistent state if BPDUs are not received.