To GVRP or not to GVRP?

[Guest]

Archived from groups: comp.dcom.lans.ethernet (More info?)

IEEE 802.1Q seems to be pretty clear on GVRP being a requirement for
VLANs (certainly to claim 802.1Q compliance) and I don't know of
another mechanism for switches to communicate VLANs registration and
configuration information. But I found this article:

http://groups-beta.google.com/group/comp.dcom.lans.ethernet/msg/ff9bcb9713b76932

which says GVRP isn't widely deployed. If VLAN requires GVRP, then
that implies VLAN isn't widely deployed. Which seems wrong. I imagine
I'm missing something. Can anyone shine some light on this for me?
Thanks.

Chris

 

[Guest]

Archived from groups: comp.dcom.lans.ethernet (More info?)

Christopher Nelson wrote:
> IEEE 802.1Q seems to be pretty clear on GVRP being a requirement for
> VLANs (certainly to claim 802.1Q compliance) and I don't know of
> another mechanism for switches to communicate VLANs registration and
> configuration information. But I found this article:
>
>
http://groups-beta.google.com/group/comp.dcom.lans.ethernet/msg/ff9bcb9713b76932
>
> which says GVRP isn't widely deployed. If VLAN requires GVRP, then
> that implies VLAN isn't widely deployed. Which seems wrong. I
imagine
> I'm missing something. Can anyone shine some light on this for me?

GVRP is a protocol used for dynamic _propagation_ of VLAN configuration
from edge (either end-station or switch) to the core. This means that
the edge must first be configured by hand. Most administrators won't
trust end-stations with configuring the network, so I would doubt that
are any significant deployments of GVRP in end-stations.

802.1Q requires implementation of GVRP to be fully conformant. Almost
all switches do implement GVRP but for some reason it hasn't been
widely deployed. Most network administrators prefer to hand-configure
VLAN membership all the way from edge to core. Use of VLANs doesn't
_require_ GVRP. GVRP was put in there to make the job of configuration
easier, but most people don't use it. It's kind of like having a
dynamic
routing protocol but choosing to use static routes instead.

Anoop

 

[Guest]

Archived from groups: comp.dcom.lans.ethernet (More info?)

sean wrote:
> anoop wrote:
>
> > 802.1Q requires implementation of GVRP to be fully conformant.
>
> I've been using (setting up, cinfiguring) vlans for about 7 years
now,
> and I had never even heard of GVRP until about 2 years ago.

It's been around since 1998 and per IEEE 802.1Q-1998, it is
mandatory.

>
> > Almost
> > all switches do implement GVRP
>
> Really? Not any of my nortel (baystack 450 and 350) or any of the
old
> cabletron switches (don't recall models) I had at my last job. Oddly
> enough one of my higer end Netgear switches does support it...

When were these switches built? I'm guessing (but don't know for a
fact) that most mid- to high-end switches built after 1998 would have
GVRP just to satisfy compliance to IEEE 802.1Q.

Anoop

 

[Sean]

Archived from groups: comp.dcom.lans.ethernet (More info?)

anoop wrote:

> 802.1Q requires implementation of GVRP to be fully conformant.

I've been using (setting up, cinfiguring) vlans for about 7 years now,
and I had never even heard of GVRP until about 2 years ago.

> Almost
> all switches do implement GVRP

Really? Not any of my nortel (baystack 450 and 350) or any of the old
cabletron switches (don't recall models) I had at my last job. Oddly
enough one of my higer end Netgear switches does support it...

> but for some reason it hasn't been
> widely deployed. Most network administrators prefer to hand-configure
> VLAN membership all the way from edge to core. Use of VLANs doesn't
> _require_ GVRP. GVRP was put in there to make the job of configuration
> easier, but most people don't use it. It's kind of like having a
> dynamic
> routing protocol but choosing to use static routes instead.
>
> Anoop
>