[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [pmtud] I am currently updating draft-ietf-pmtud-method-02bis
Nothing so complicated. I am looking for a table of likely tunnel header
overheads (sizes). If you know that the base MTU is 1500, what are the likely
useful MTUs across all types of tunnel environments? I don't care if the
tunnel endpoints are end systems or middle boxes, or any of the other details.
This is for the PLPMTUD search heuristic.
Let me give a specific question: If a 1044 Byte packet is delivered, and a
1500 Byte packet is not, what other sizes would you try, and why?
Is is sufficient to assume multiples of 12, 16 or 20 bytes? Which?
Or would it be significantly to have an explicit table of common tunnel sizes?
Thanks,
--MM--
-------------------------------------------
Matt Mathis http://www.psc.edu/~mathis
Work:412.268.3319 Home/Cell:412.654.7529
-------------------------------------------
Evil is defined by people who think they know
"The Truth" and use force to apply it to others.
On Wed, 20 Oct 2004, Fred Templin wrote:
Matt Mathis <mathis at psc.edu> wrote: Quick question: can somebody cite a table of common tunnel overheads?
For example:
20 : IPv4 in IP
12 : IPsec ESP_AUTH
24 : IPsec AH, IPsec ESP_CIPHER
Matt,
I'm wondering about the nature of your concern here . Are you worried about
the overhead for encapsulations that occur between the packetization layer
and the link layer within the same end node that performs packetization, i.e.,
"transport-mode" IPsec/VPN/etc.? Or, are you concerned about "tunnel mode"
encapsulations that might occur in middle boxes along the path? (Maybe my
terminology is off here, but I am assuming you meant the former even though
you said "tunnel overheads" - please correct me if I'm wrong.
Assuming you are talking about encapsulations that get added between the
packetization layer and the link layer, I suppose there are two components
to the question: 1) how does the packetization layer know which paths will
add the encapsulations so that it can reduce the MPS for those paths, and
2) how does the packitization layer deal with ICMP PTB messages that
report a larger MTU size than the actual size of the packetization layer
segment used for probing (larger because IPsec/VPN/etc encapsulations
were added after packetization)?
Thanks - Fred
osprey67 at yahoo.com
_______________________________________________
pmtud mailing list
pmtud at ietf.org
https://www1.ietf.org/mailman/listinfo/pmtud