25. Spanning Tree
ROX™ v2.2 User Guide
234
RuggedBackbone™ RX5000
25.1.2. Edge Ports
A port may be designated an Edge Port if it is directly connected to an end station. As such, it cannot
create bridging loops in the network and can thus directly transition to forwarding, skipping the listening
and learning stages.
Edge ports that receive configuration messages immediately lose their Edge Port status and become
normal spanning tree ports. A loop created on an improperly connected edge port is thus quickly
repaired.
Because an Edge Port services only end stations, topology change messages are not generated when
its link toggles.
25.1.3. Point-to-Point and Multipoint Links
RSTP uses a peer-peer protocol called Proposing-Agreeing to ensure transitioning in the event of a link
failure. This protocol is point-to-point and breaks down in multipoint situations, i.e. when more than two
bridges operate on a shared media link.
If RSTP detects this circumstance (based upon the port’s half duplex state after link up) it will switch
off Proposing-Agreeing. The port must transition through the learning and forwarding states, spending
one forward delay in each state.
There are circumstances in which RSTP will make an incorrect decision about the point-to-point state
of the link simply by examining the half-duplex status, namely:
• The port attaches only to a single partner, but through a half-duplex link.
• The port attaches to a shared media hub through a full-duplex link. The shared media link attaches
to more than one RSTP enabled bridge.
In such cases, the user may configure the bridge to override the half-duplex determination mechanism
and force the link to be treated in the proper fashion.
25.1.4. Path and Port Costs
The STP path cost is the main metric by which root and designated ports are chosen
1
. The path cost
for a designated bridge is the sum of the individual port costs of the links between the root bridge and
that designated bridge. The port with the lowest path cost is the best route to the root bridge and is
chosen as the root port.
How Port Costs Are Generated
Port costs can be generated either as a result of link auto-negotiation or manual configuration.
When the link auto-negotiation method is used, the port cost is derived from the speed of the link.
This method is useful when a well-connected network has been established. It can be used when the
designer is not too concerned with the resultant topology as long as connectivity is assured.
Manual configuration is useful when the exact topology of the network must be predictable under all
circumstances. The path cost can be used to establish the topology of the network exactly as the
designer intends.
1
In actuality the primary determinant for root port selection is the root bridge ID. Bridge ID is important mainly at network startup when
the bridge with the lowest ID is elected as the root bridge. After startup (when all bridges agree on the root bridge’s ID) the path cost
is used to select root ports. If the path costs of candidates for the root port are the same, the ID of the peer bridge is used to select
the port. Finally, if candidate root ports have the same path cost and peer bridge ID, the port ID of the peer bridge is used to select
the root port. In all cases the lower ID, path cost or port ID is selected as the best.
