From: John Gibson (johngibson1541@yahoo.com)
Date: Tue Apr 24 2007 - 18:25:39 ART
I thought we don't need to increase mtu in 3550 or
3560 in order to run dot1q or isl. Maybe I am wrong.
I thought "baby giant" is supported everywhere.
univercd is saying "1504" to run dot1q tunneling.
How could 1504 be enough to run ISL. Any trunk takes
more than 4 bytes. Unless the service provider doesn't
run trunk.
I am totally lost.
John
--- Robert Hosford <rhosford@certifiednets.com> wrote:
> John,
>
> Personally I wish that every device would support
> Jumbo frames. There are
> so many headers that I would like to be supported so
> we don't have to
> fragment as much as we do. ISL, Dot1.q, IPSEC, GRE,
> MPLS to name a few
> places frame growth should be allowed.
>
>
> Robert
> -----Original Message-----
> From: nobody@groupstudy.com
> [mailto:nobody@groupstudy.com] On Behalf Of John
> Gibson
> Sent: Tuesday, April 24, 2007 1:46 PM
> To: ccielab@groupstudy.com
> Subject: dot1q tunneling "system mtu" is 3550, 3560
> only and should be
> considered "caveats" ?
>
> We know span destination reflector port is 3550 only
> and should be considered "caveats".
>
> I am looking at 6500 switch's document , it doesn't
> say any thing about increasing mtu by 4 bytes for
> dot1q tunneling.
> 6500 switch's document generalizes the issue as
> "jumbo frame support".
>
> Every switch's software should only send packets at
> 1500 maximum but allow receiving a lot more bytes
> as a measure of tolerance right ? When we do some
> protocol, we don't assume we know what other unknown
> protocols are doing to our packets right ? We don't
> know what we don't know right ?
>
> The increasing 4 byte thing will disappear soon
> just like as soon as 3550 is gone, span reflector
> port is gone right ?
>
>
>
> __________________________________________________
> Do You Yahoo!?
> Tired of spam? Yahoo! Mail has the best spam
> protection around
> http://mail.yahoo.com
>
>
This archive was generated by hypermail 2.1.4 : Tue May 01 2007 - 08:28:37 ART