This read-only counter tallies frames discarded for any reason on this connection, including those described elsewhere in this table as well as causes such as insufficient buffer space during reassembly. Because this is an explicit superset that also captures causes not individually broken out, it should always be at least as large as the sum of the reassembly-timeout, length-error, and CRC-error counters, and a gap between this total and those specific counters points to a discard cause, most likely buffer exhaustion, that isn't otherwise being individually tracked. This total running noticeably higher than the sum of the reassembly-timeout, length-error, and CRC-error counts combined on a heavily loaded connection would indicate the switch is running short on reassembly buffers for that VC.