CFT: new trunk(4)

[Lars Erik Gullerud]

On Thu, 12 Apr 2007, Ian FREISLICH wrote:

> Peter Jeremy wrote:
>>
>> 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'm not going to tell you that, I am however going to give you the 802.3ad 
(Link Aggregation) explanation.

Quote from IEEE 802.3 2005 edition, clause 43.1:

"Link Aggregation allows one or more links to be aggregated together to 
form a Link Aggregation Group, such that a MAC Client can treat the Link 
Aggregation Group as if it were a single link. To this end, it specifies 
the establishment of DTE to DTE logical links, _consisting of N parallell 
instances of full duplex point-to-point links operating at the same data 
rate._"

VLANs are not MAC Clients in 802 terminology, for the 802.3 MAC sublayer, 
VLANs are opaque. VLAN "interfaces" in FreBSD are purely logical 
entities that does not have MAC sublayer as such. (The MAC sublayer, as 
opposed to the LLC sublayer, deals mostly with low-level communications 
issues such as Auto-neg and CSMA/CD)

What you want to do is possible using other proprietary methods (or 
better yet, on a different OSI Layer - sounds more like you 
want layer 3 load balancing), but it can not be done with Link 
Aggregation, which the oh-so-badly-named trunk(4) interface implements. 
This is not a FreeBSD issue.

And yes, as another poster has mentioned - when one speaks of a trunk in 
Ethernet terminology, one generally refers to a link carrying multiple 
tagged VLANs.

/leg