A number of bridge vendors offer proprietary features whereby two separate
bridges appear to all external devices (including the other bridges in the network) to
be a single bridge. Multiple connections between a single bridge and the coupled pair,
or between two coupled pairs, can be made using link aggregation. If one bridge of the
coupled pair fails, the other takes over, with the disruption confined to the inevitable
loss of the bandwidth in the links connected to the failed bridge; the spanning tree
topology need not change at all. (The system administrator may elect to allow the topology
to change quickly, in this situation, to account for the lost bandwidth.)
Failed Bridges As mentioned previously in ???Spanning Tree Convergence,??? when the
spanning tree software process fails in a bridge, in a manner that does not bring down
the whole bridge, a permanent forwarding loop can result that can severely disrupt the
network operation. Fortunately, this problem is solvable. The first obvious solution is,
of course, to build more reliable software, or at least software that is able to detect its
failures and recover from them more readily and thus be more robust.
Pages:
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906