Portfast, Uplinkfast, Backbonefast
examseeker
Member Posts: 118
in CCNP
Ok.. have you ever had the feeling where your eyes start crossing on certain topics? I am trying to get a good feeling of the topics of this post. Could you provide a little 'lesson' for me? Just how do these work? I understand Portfast because I've read it at least twice.
Pros and cons of the above (in addition to an explanation)?
Thanks -- you have my graditude..
es
Pros and cons of the above (in addition to an explanation)?
Thanks -- you have my graditude..
es
Comments
-
nevolved Member Posts: 131You don't know these as a CCNA? Those are part of the objectives for ICND2.
-
aordal Member Posts: 372Someone correct me if I'm wrong because this is my understanding.
Portfast tells a port not to take place in STP, you would set this on a port(s) that do not connect to switches. Set this on ports that connect directly to PCs.
Backbone fast you would not set on a Access Switch (switches that have PCs plugged into them) but the Distribution Layer switch. This setting assists remote switches in switching to another redundant link(trunk) in the event that the current link(trunk) it is currently using goes down.
Uplink fast you would set on an Access Switch (not a Distribution Switch) and it only works locally on itself to assist in switching to a 2nd redundant link(trunk) in the event that the current link(trunk) fails. -
examseeker Member Posts: 118I know them - I just wanted more clarification.. the Cisco site has loads of details..
Thanks -
kpjungle Member Posts: 426Portfast is set to enable host's to get on the network immediately, and therefore bypassing the normal STP (802.3d) transition from blocking->listening->learning->forwarding. It sets it immediately to forwarding state.
Uplinkfast is the traditional (802.3d) way of having a port in "ready-to-take-over" state. If an existing uplink fails, this one can take over immediately. Somewhat similar to Rapid stp's backup port role.
Backbonefast is meant to resolve a connection failure on the backbone. By using this, if a switch receives a root-bpdu with the sending's switch's address as the root, it will know it has a broken link. The receiving switch can then send out a RLQ (root link query), to see if it still has a root itself, if it gets a response to this, it can inform the "offending" switch, and the transition will occur faster than the listening + learning times (30 seconds by default).
Dont know if that clarified anything thoughStudying for CCNP (All done)