Thus spake "Nimesh Vakharia" <nvakhari@clio.rad.sunysb.edu>
> Well but I think there have been a lot of enhancements after that like
> PVST i.e. Per Vlan Spanning Tree. i've seen instances where a vlan trunk
> is blocking for one vlan but active for another depending on the topology.
> In this scenario BPDU's are local for that vlan broadcast domain
> indicating they pare tagged.
PVST existed long before 802.1Q did. In fact, the failure to include PVST
in 802.1Q was because some vendors didn't have the CPU power to or coding
expertise to run multiple STP's. Recent work on MST (802.1s) addresses this
shortcoming.
> No other vendor seems to have a concept of native vlans but Cisco, as
> soon as you tag/untag a vlan to a port in fdry or extreme you lose all
> the default vlan associations to it.
The default VLAN for other vendors means the same thing as Cisco's native
VLAN, just a different term. It's the VLAN frames are assumed to be on if
they are not tagged. Many protocols, such as STP (non-PVST) are specified
to always be untagged.
I can't speak to exactly how Foundry or Extreme handle the default VLAN on a
port, or if they even allow you to access it. I strongly recommend to
customers they not use a native/default VLAN for user traffic anyways,
specifically due to security and stability concerns.
> Well auto neg dont' dosen't really communicate via ethernet
> frames so its independent of vlans but i think i see what you are
> saying, Cisco proprietary DTP could probably make of the Native Vlans.
Any control traffic that is unrelated to a particular VLAN will be untagged,
eg. CDP, PAgP, 802.1x, etc.
S
This archive was generated by hypermail 2b29 : Sun Aug 04 2002 - 04:13:27 EDT