Link Failure Handling

When a link in the ring goes down, the ring splits into two separate cascades. This occurs in the following sequence of events:
  1. Link Failure Detection and the Ring Link Fail Notification (RLFN) Message
  2. Ring Disable Message
  3. Traffic Re-routing and Cascade Information Clearing
  4. Failed Link Restoration

When the connection between two BPEs in a ring is severed, the BPE whose upstream port was at the point of disconnection sends a notification of the failure to its active cascade port. Simultaneously, that BPE will "fail over" to its backup upstream port, changing the upstream direction of the data plane. The notification is called a Ring Link Failure Notification (RLFN). When the downstream BPE receives it, it forwards the RLFN to its active cascade port. If the BPE received the RLFN over its active upstream port, it too fails over as above. The end result is one or two data plane cascades terminating at the broken connection with no remaining backup path, which is re-established after the next ring detection. The new backup upstream direction is established on all nodes. Some BPEs will have different upstream and backup upstream ports than those that were originally established. The cascades that were established when the ring was severed remain in place.