RE: email, smtp ( was Re: bypass requirement?)

["Stecher,Martin" <martin_stecher@xxxxxxxxxxxxxxxxxxx>]

To clarify:
The "MUST" is a requirement for the OPES/SMTP specs draft to provide a
bypass feature. Whether the OPES/SMTP specs then requires that this feature
MUST be implemented is a different topic.

OPES itself is application agnostic.
Different adaptation documents can be written to handle specifics of certain
protocols.
Current work concentrates on SMTP. You could go and do something for POP3 or
IMAP :-)

Martin

> -----Original Message-----
> From: ArkanoiD [mailto:ark@xxxxxxxxx] 
> Sent: Mittwoch, 13. Dezember 2006 11:39
> To: Stecher,Martin
> Cc: ietf-openproxy@xxxxxxx
> Subject: email, smtp ( was Re: bypass requirement?)
> 
> nuqneH,
> 
> I think "must" word is evil sometimes. IMAP comes in mind, 
> the protocol that seems to be designed to make content 
> inspection a real nightmare - just because every way to slice 
> a email to tiny bits is mandatory to be implemented on the server.
> 
> Back to OPES, shouldn't there be "generic OPES for email", 
> not really smtp-specific but able to deal with SMTP, POP3 and 
> (bwarf!) IMAP?
> 
> On Wed, Dec 13, 2006 at 10:11:39AM +0100, Stecher,Martin wrote:
> > 
> > 
> > OPES specificationas are not created for single environments. 
> > While a bypass feature may not be useful in one scenario it will be 
> > very valid in others.
> > 
> > Following earlier discussions and IAB requirements, OPES specs MUST 
> > provide a bypass feature while the local policy when deploying and 
> > implementing OPES in certain environments will determine whether a 
> > bypass request can be fulfilled or not.
> > 
> > Martin
> > 
> > > -----Original Message-----
> > > From: owner-ietf-openproxy@xxxxxxxxxxxx 
> > > [mailto:owner-ietf-openproxy@xxxxxxxxxxxx] On Behalf Of ArkanoiD
> > > Sent: Dienstag, 12. Dezember 2006 16:10
> > > To: ietf-openproxy@xxxxxxx
> > > Subject: bypass requirement?
> > > 
> > > 
> > > nuqneH,
> > > 
> > > Doesn't mandatory OPES bypass feature requirement (it is noted as 
> > > "must") look illogical in firewall environment?
> > > 
> > > First, it makes code more complicated and second, it 
> should be never 
> > > used ;-)
> > > 
> > > 
> 
>