CFT: new trunk(4)

Fri Apr 13 16:49:30 UTC 2007

Andrew Thompson wrote:
> On Thu, Apr 12, 2007 at 07:39:00AM +0200, Ian FREISLICH wrote:
> 
>>Peter Jeremy wrote:
>>
>>>On 2007-Apr-11 15:43:04 +0200, Ian FREISLICH <ianf at clue.co.za> wrote:
>>>
>>>>Andrew Thompson wrote:
>>>>
>>>>>On Mon, Apr 02, 2007 at 11:17:29AM +0200, Ian FREISLICH wrote:
>>>>>
>>>>>>We're making extensive use of vlans to increase the number of
>>>>>>interfaces availabble to us using switches to break out gigE into
>>>>>>100M interfaces.  The bandwidth problem we're having is to our
>>>>>>provider, a 100M connection, and we're looking at doing exactly
>>>>>>this.  However, it appears that this interface can't trunk vlan
>>>>>>interfaces.
>>>
>>>=2E..
>>>
>>>>No, I'm sure I want it the way I said.  I know it sounds wrong, but
>>>>I just don't have enough PCI-X slots to waste 2 on physical 100M
>>>>NICs for the uplink from the routers.
>>>
>>>Trunking is a way of combining multiple physical interfaces to increase
>>>the bandwidth.  Trunking multiple VLANs on a single interface doesn't
>>>make sense to me.
>>
>>802.1q is VLAN tagging and trunking.  This interface is LACP - link
>>aggregation.  I really think that it makes no sense to be able to
>>aggregate some ethernet interfaces and not others.  I suppose some
>>pedant will tell me vlan interfaces are not ethernet.
> 
> 
> I think the unfortunate name of trunk(4) that we inherited from OpenBSD
> is causing quite some confusion.  trunk(4) actually has nothing to do
> with vlan trunking which I think you are after.
> 
> I can see this topic coming up again so it could save some time to
> rename the driver now. It would mean that we lose the naming link to the
> same driver in OpenBSD but you cant win em all.
> 
> Some names that have been suggested are:
> 
> linkag(4)
> agr(4)
> bond(4)   <- same as linux
> 
> Any suggestions!

lacp(4)++

-- 
Andre