Re: [secdir] [Isms] secdir review ofdraft-ietf-isms-transport-security-model-12

["David B Harrington" <[e-mail]>]

OK, I'll change my answer. 

That is a deployment decision made by an administrator who has an
understanding of what is appropriate to the system in question.

What is the correct non-RFC2119 phrase in which to couch our
deployment advice?

dbh

> -----Original Message-----
> From: Jeffrey Hutzelman [mailto:[e-mail]] 
> Sent: Tuesday, May 05, 2009 3:10 PM
> To: David B Harrington; 'Barry Leiba'; [e-mail];
[e-mail]
> Cc: [e-mail]; 
> [e-mail]; [e-mail]; [e-mail]
> Subject: Re: [Isms] [secdir] secdir review 
> ofdraft-ietf-isms-transport-security-model-12
> 
> --On Tuesday, May 05, 2009 01:50:46 PM -0400 David B Harrington 
> <[e-mail]> wrote:
> 
> > That is a deployment decision made by an administrator who has an
> > understanding of what is appropriate to the system in question.
MUST
> > is for implementers and SHOULD is for deployers.
> 
> Nononono.  When we say "MUST is for implementors", we mean that our 
> specifications give requirements for implementations which 
> claim to comply; 
> RFC2119 requirements language is generally inappropriate for 
> deployment 
> advice.
> 
> It is appropriate to say TSM MUST be use with a TM that provides 
> appropriate security; that is equivalent to saying that 
> implementations 
> MUST NOT use TSM together with an inappropriate TM, such as UDP.
> 
> 
> That said, selection of transport and security models generally is
an 
> operational decision.  TSM itself is really architectural 
> glue to allow us 
> to push some of the security infrastructure down into a 
> transport below 
> SNMP, such as SSH or TLS.  You'd never say "I want to use 
> TSM, now which 
> transport should I use"; you'd say "I want to use SSH -- oh, 
> for that to 
> work, I have to use it with TSM".
> 
> 
> -- Jeff
>