EXP Computer S200 Manuel D’Utilisation

Bridging 

97

T0008-16F

Release 5.2M

For transparent bridges the bridge Forward Delay is used to allow the spanning tree 
algorithm to converge to a stable topology before the bridging process is allowed to 
proceed. Spanning tree topology determination is an iterative process and requires 
time to converge. The bridge should not forward packets during this time because 
temporary loops might cause forwarded packets to be exponentially duplicated and 
disable the network. Once the topology has stabilized, the bridge should not forward 
packets immediately since, initially, it will not have learned station locations and will 
have to broadcast packets when it does not find the entry in its local station cache. 
Once the bridge has built up its cache by listening for a short period of time, it can 
forward packets directly, rather than using high overhead broadcasts.

The Forward Delay is used twice: once to allow the topology to stabilize and during 
this time to process only spanning tree protocol messages; and then again to allow 
the bridge to learn station locations, during which time data packets are received but 
not forwarded.

To understand how long it takes to converge a spanning tree, consider the following 
simplified network shown in Figure 47, together with the message events shown as a 
timed sequence below the network:

Bridge 1

Bridge 3

Bridge 2

Bridge 4

Bridge IDs

Link Numbers