Configuring Edge Safeguard

Loop prevention and detection on an edge port configured for RSTP is called edge safeguard. You can configure edge safeguard on RSTP edge ports to prevent accidental or deliberate misconfigurations (loops) resulting from connecting two edge ports together or by connecting a hub or other non-STP switch to an edge port. Edge safeguard also limits the impact of broadcast storms that might occur on edge ports. This advanced loop prevention mechanism improves network resiliency but does not interfere with the rapid convergence of edge ports.

An edge port configured with edge safeguard immediately enters the forwarding state and transmits BPDUs. If a loop is detected, STP blocks the port. By default, an edge port without edge safeguard configured immediately enters the forwarding state but does not transmit BPDUs unless a BPDU is received by that edge port.

You can also configure edge safeguard for loop prevention and detection on an MSTP (Multiple Spanning Tree Protocol) edge port.

In ExtremeXOS 11.5 and earlier, ports that connect to non-STP devices are edge ports. Edge ports do not participate in RSTP, and their role is not confirmed. Edge ports immediately enter the forwarding state unless the port receives a BPDU. In that case, edge ports enter the blocking state. The edge port remains in the blocking state until it stops receiving BPDUs and the message age timer expires.

ExtremeXOS 11.6 and later support an enhanced bridge detection method, which is part of the 802.1D-2004 standard. Ports that connect to non-STP devices are still considered edge ports. However, if you have an 802.1D-2004 compliant edge port, the bridge detection mechanism causes the edge port to transition to a non-edge port upon receiving a BPDU. If the former edge port does not receive a subsequent BPDU during a pre-determined interval, the port attempts to become an edge port.

In ExtremeXOS 12.0.3 and 12.1.4 onwards, STP edge safeguard disables a port when a remote loop is detected. ExtremeXOS versions prior to 12.0.3 and 12.1.4 place the port in blocking mode. The change was made because BPDUs are still processed when a port is in a blocking state. A remote loop causes BPDUs to be exponentially duplicated which caused high CPU utilization on the switch even though the port was transitioned to a blocked state.