Re: [midcom] security recommendations in MIDCOM MIB draft

[Melinda Shore <mshore at cisco.com>]

On 7/12/07 7:51 AM, "Magnus Westerlund" <magnus.westerlund at ericsson.com>
wrote:
> Can we please come to consensus on this topic. And if there are text
> changes to implement the consensus, please provide them as RFC-editor
> notes to me.

The starting point is: requesting services from a middlebox must
be secure.  If that's to be done cryptographically, it requires
SNMPv3.  If it's not to be done cryptographically it suggests that
the protocol is being run over a "secure" network.  The latter was
not considered acceptable by the responsible area director at the
time that the work was ramping up, and it seems to me the question
of whether or not SNMPv3 is to be required hinges on whether or
not we can now permit an assumption of a "secure" network (for example,
running it over an IPSec tunnel).  It seems to me that if we
can't assume the use of a secure network in some deployments then
SNMPv3 has to be required, since sending firewall requests and
NAT answers insecurely is obviously unacceptable.  Do you think
the assumption of a secure network would pass review?

Melinda

_______________________________________________
midcom mailing list
midcom at ietf.org
https://www1.ietf.org/mailman/listinfo/midcom