From what i have read from the book, can i say that the placing of the following feature are correct from the below answer stated:

Portfast:
Only access switches where interface invlove end-user ports, cisco devices and server ports.

Uplinkfast:
Only access switches where the number of active VLAN is limited.

BackboneFast:
Between Distribution and Core switches where link cores encounter direct failures.

BPDU Guard:
Only access switches where interface configure with portfast feature.

BPDU Filtering:
Only access switches where interface configure with portfast feature.

Root Guard:
Between distribution and access switches.

Loop Guard:
Between Core, Distribution and Core switches where link cores encounter indirect failures.

UDLD:
All Core, Distribution and Core switches where link cores encounter physically link failures.

Moved to Tech Discussions.

_________________
Reasonably un-nerdy blog:
americanwerewolfinbelgrade.wordpress.com/

Never use this !!

Can I ask why? I am just now deploying a fix that involves this command.

_________________
Find networking-forum.com on Facebook, LinkedIn, Twitter, Google+,or subscribe to the site's RSS feeds.

It disables spanning tree completely on a port that you apply it to. It supersedes BPDU guard, so even if that port received a BPDU, it would ignore it instead of shutting down the port. If someone accidentally plugged a switch into it, you would have a loop and get a broadcast storm. I learned this the hard way. I only use two commands now for access ports-

spanning-tree portfast default
spanning portfast bpduguard default

The first enables portfast globally on all access ports and the second enables bpduguard on all portfast ports. So the access ports will still send out bpdu's, but if someone plugs a switch into that port or if someone maliciously sends a bpdu into that port, it will go into err disable state.

I do use BPDU filter in a lab environment if I want to deliberately create a broadcast storm. Last time I was using it to ramp up the CPU utilization to see if an SNMP system was reading the processor correctly. It was reading it good alright. I had the CPU at 97% from a few DHCP requests.