Wireshark · Ethereal-users: Re: [Ethereal-users] (Slightly OT) Ethernet length question

[Jaap Keuter <jaap.keuter@xxxxxxxxx>]

Hi,

You skipped the a) clause! This is from the 1998 version I've got laying
around:

7.2 Support of the Internal Sublayer Service by IEEE Std 802.3 (CSMA/CD)
......
a) Use the procedure as described in ISO/IEC 15802-3, 6.5.1. This can
result in tagged frames of less than 68 octets (but at least 64 octets)
being transmitted; or
......

So at least 64 octets on the wire, either with or without VLAN tag.

Sorry,
Jaap


On Wed, 18 Jan 2006 Ran.Shenhar@xxxxxxxxxxx wrote:

> OK, thanks for the clarification.
> However, in the meantime I also conducted another search, and I found two
> sources -
> 1. IEEE 802.1Q (2003), see pg 39 (actual page, including all prefaces etc),
> look near the end (specifically at the sentence "Include additional octets
> before the FCS field in order for the transmitted frame length for such
> frames to be 68 octets. This results in a minimum tagged frame length of 68
> octets.")
> (the file can be found here: ftp://ftp.ecitele.com/notes/IEEE802.1-Q.-2003
> _(VLAN).pdf)
> 2. According to CISCO - 68 bytes (
> http://www.cisco.com/en/US/tech/tk389/tk390/technologies_tech_note09186a0080094665.shtml#frame2
> )
>
> Now, my problem is that I got a bet with R&D guys around here, and it seems
> my sources say that it's 68 bytes, and you're saying it's 64... or have I
> misunderstood the IEEE doc??
>
> TnX
>
>
>
>              Jaap Keuter
>              <jaap.keuter@xs4a
>              ll.nl>                                                     To
>              Sent by:                  Ethereal user support
>              ethereal-users-bo         <ethereal-users@xxxxxxxxxxxx>
>              unces@xxxxxxxxxxx                                          cc
>              m
>                                                                    Subject
>                                        Re: [Ethereal-users] (Slightly OT)
>              18/01/2006 13:34          Ethernet length question
>
>
>              Please respond to
>                Ethereal user
>                   support
>              <ethereal-users@e
>                thereal.com>
>
>
>
>
>
>
> Hi,
>
> What you should decouple is the minumum 64 octet Ethernet requirement
> from the actual use of these octets.
>
> The 802.1Q spec only tells you that if the Ethernet II header contains a
> type value of 0x8100 the next two octets are to be interpreted as VLAN
> tag. That defines use.
>
> When a frame is present on the wire it needs to be minimum 64 octets,
> whatever the use of these octets.
>
> These are seperate things.
>
> Hope it helps,
> Jaap
>
>
>
> On Wed, 18 Jan 2006 Ran.Shenhar@xxxxxxxxxxx wrote:
>
> > TnX Jaap,
> > I know the CSMA/CD principles behind the 64 byte minimum requirement.
> > However, my question is really focused on the 802.1Q change (or no
> change)
> > in that demand, and I didn't understand the answer for that issue from
> your
> > answer.
> > Could you please try to clarify?
> >
> > TnX
> >
> >
> >
> >              Jaap Keuter
> >              <jaap.keuter@xs4a
> >              ll.nl>
> To
> >              Sent by:                  Ethereal user support
> >              ethereal-users-bo         <ethereal-users@xxxxxxxxxxxx>
> >              unces@xxxxxxxxxxx
> cc
> >              m
> >
> Subject
> >                                        Re: [Ethereal-users] (Slightly OT)
> >              18/01/2006 11:54          Ethernet length question
> >
> >
> >              Please respond to
> >                Ethereal user
> >                   support
> >              <ethereal-users@e
> >                thereal.com>
> >
> >
> >
> >
> >
> >
> > Hi,
> >
> > The layer protocol model is the best reference you can take in mind.
> > The minimum ethernet frame length (that is bits on the wire) is 64*8.
> > Together with the wirespeed this gives a minimum amount of transmission
> > time. This time is needed for the other endpoint (in a point-to-point
> > link) or other stations (on a shared medium) to assure recognition of the
> > frame and collision avoidance. All this is Physical layer stuff.
> >
> > Then comes the meaning of the bits in the frame. If the ethernet header
> > says that it's a VLAN tag that is following then the next 2 octets are to
> > be interpreted that way. From there on the next protocol layer starts.
> >
> > That is the true meaning of VLAN, a VIRTUAL LAN. It's not a real one, as
> > seen on the wires, but on a layer above that.
> >
> > Hope it helps,
> > Jaap
> >
> >
> > On Wed, 18 Jan 2006 Ran.Shenhar@xxxxxxxxxxx wrote:
> >
> > > Hi All,
> > > I need help with a non-ethereal specific - I know that an ethernet
> frame
> > is
> > > of minimum length = 64 bytes, including FCS.
> > > However, what is the minimum length of a VLAN tagged ethernet frame? Is
> > it
> > > 64+4, or is it still 64?
> > >
> > > I tried looking in IEEE 802.3q 2003 edition, however I was not able to
> > find
> > > a definite answer.
> > > If you could also point me to a reference about the answer, it'd be
> even
> > > better.
> > >
> >
> > _______________________________________________
> > Ethereal-users mailing list
> > Ethereal-users@xxxxxxxxxxxx
> > http://www.ethereal.com/mailman/listinfo/ethereal-users
> >
> >
> > _______________________________________________
> > Ethereal-users mailing list
> > Ethereal-users@xxxxxxxxxxxx
> > http://www.ethereal.com/mailman/listinfo/ethereal-users
> >
>
> _______________________________________________
> Ethereal-users mailing list
> Ethereal-users@xxxxxxxxxxxx
> http://www.ethereal.com/mailman/listinfo/ethereal-users
>
>
> _______________________________________________
> Ethereal-users mailing list
> Ethereal-users@xxxxxxxxxxxx
> http://www.ethereal.com/mailman/listinfo/ethereal-users
>