Re: [dnsext] EDNS Page Option to handle large responses

[David Conrad <[e-mail]>]

On Jun 15, 2009, at 7:00 PM, Patrik Fältström wrote:
> On 16 jun 2009, at 00.47, George Barwood wrote:
>
>> EDNS responses larger than ~1500 bytes currently suffer from 2  
>> problems due to IP fragmentation :
>> (1) They may fail to arrive due to IP implementation or equipment  
>> configuration issues.
>
> I must say that implementing application layer "features" to solve a  
> problem when IP layer is not behaving properly (can not handle  
> fragmentation) is a layering violation in the architecture.

It just means that you (typically) need to replicate logic found in  
connection oriented transport protocols.  See (original) NFS for an  
example.  I guess you could call it a layering violation, but if so,  
allowing applications to choose between connection-oriented and  
connectionless would be the same sort of layering violation.

> If we do have problems with fragmented packets not being  
> reassembled, then I think we have other problems as well in our  
> networks.


?

http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.37.5308
http://tools.ietf.org/html/draft-heffner-frag-harmful-05
The whole point behind RFC 1191
Etc.

Regards,
-drc


--
to unsubscribe send a message to [e-mail] with
the word 'unsubscribe' in a single line as the message text body.
archive: <http://ops.ietf.org/lists/namedroppers/>