[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Sigtran] Clarify SCTP max packet size.



Brian,

I'm trying to figure out what my mistake is...

Are you saying that the SCTP layer will construct SCTP
messages which needs to be fragmented at the IP layer?
(not considering the case where the path MTU drops
and retransmissions are required)

I think Pierre assumed that SCTP constructs SCTP messages
of 1500 bytes which have no space anymore for the IP
header when sent over an 1500 byte MTU ethernet interface.
And I do not think that this is the case. Am I wrong?

Best regards
Michael

On Jul 8, 2009, at 1:19 AM, Brian F. G. Bidulock wrote:

Michael,

I believe I was correcting your statement.

--brian

Michael Tüxen wrote: (Wed, 08 Jul 2009 00:43:33)
Hi Brian,

sorry, I do not understand what you want to say...

Best regards
Michael

On Jul 7, 2009, at 8:02 PM, Brian F. G. Bidulock wrote:

Michael,

Err, the MTU SCTP uses when bundling is the Association Path MTU. It subtracts IP header and SCTP header sizes from this Path MTU only when
considering how many chunks can be bundled.

--brian

Michael Tüxen wrote:                        (Tue, 07 Jul 2009
18:51:09)
On Jul 7, 2009, at 8:50 AM, <pierre.allegret at orange-ftgroup.com>
<pierre.allegret at orange-ftgroup.com > wrote:

Dear all,

Just to add with the SCTP bundling feature. If the "SCTP MTU" is
1500, and the system is bundling messages at the SCTP level, you
should have the risk to segment at the IP level (because of the IP
header). So it is not efficient
The MTU SCTP uses is the link layer MTU minus the headers for IP and
SCTP.
I have not seen an implementation which bundles at the SCTP layer and
than
needs to fragment at the IP layer. I would consider this broken...
to bundle sctp messages and to segment them at the IP level. Some
vendors limits "SCTP MTU" to avoid that. This is only the Maximum
transmission unit. If one SCTP message to transmit is greater, they
will segment it.


--
Brian F. G. Bidulock
bidulock at openss7.org
http://www.openss7.org/



--
Brian F. G. Bidulock
bidulock at openss7.org
http://www.openss7.org/