I concur with Randy and Juergen.
-- 
Regards,
Uri Blumenthal




On 6/23/15, 10:05 , "Juergen Schoenwaelder"
<[email protected]> wrote:

>On Mon, Jun 22, 2015 at 04:12:45PM -0400, Kathleen Moriarty wrote:
>> On Mon, Jun 22, 2015 at 4:04 PM, Randy Presuhn
>><[email protected]
>> > wrote:
>> 
>> > Hi -
>> >
>> > >From: Kathleen Moriarty <[email protected]>
>> > >Sent: Jun 22, 2015 12:45 PM
>> > >To: Michael MacFaden <[email protected]>
>> > >Cc: "[email protected]" <
>> > [email protected]>, Randy Presuhn <
>> > [email protected]>, "[email protected]" <
>> > [email protected]>, "[email protected]" <
>> > [email protected]>, The IESG <[email protected]>, "[email protected]" <
>> > [email protected]>, "[email protected]" <
>> > [email protected]>
>> > >Subject: Re: [OPSAWG] Kathleen Moriarty's No Objection on
>> > draft-ietf-opsawg-vmm-mib-03: (with COMMENT)
>> > >
>> > >Hi Mike,
>> > >
>> > >Sent from my iPhone
>> > >
>> > >> On Jun 22, 2015, at 9:08 PM, Michael MacFaden <[email protected]>
>>wrote:
>> > >>
>> > >> Agree there should be some warning here. But if this text is
>>intended
>> > for those who implement it
>> > >> (like me) then I'd prefer to  make it clear as mud what the issue
>>is:
>> > >>
>> > >>  Any implementation of this MIB module in an agent where the
>>Virtual
>> > Machines being monitored
>> > >>  have access to this very agent or this MIB module creates an
>>attack
>> > vector on the system or on
>> > >>  any other VM hosted by this system.
>> > >
>> > >Yes and one of the responses to the SecDir review said this mib would
>> > just be on the hypervisor.  It also said the virtual images managed
>>by the
>> > hypervisor typically does not have access to the management
>>connection for
>> > SNMP to the hypervisor, limiting the risk.
>> > >
>> > >Shouldn't the text include guidance that makes this clear (where
>>this mib
>> > is and the reality of whether or not it's a risk from managed virtual
>> > images)?
>> > >
>> > >Recommending a separate management connection is one way to do that.
>> >
>> > I'm missing something here.  Why is a monitored VM (where the
>> > MIB instrumentation is external to the VM) different from any other
>> > threat?  It seems to me that "access to the management connection"
>> > is a red herring - the privacy protocol, along with authentication
>> > and access control (VACM) should just be left to do their job.
>> >
>> 
>> This came up in the SecDir review.  I was happy with the response
>>provided
>> via email in how it was put out-of-scope.  The draft itself does not
>>have
>> the explanation offered and I think it would be helpful to the reader in
>> case the same question comes up later.  I thought it was a reasonable
>> question given the content of the draft and possible threat space.
>>
>
>Hi,
>
>the only meaningful way to implement this MIB is as part of the
>hypervisor. I think this actually follows directly from section 3. I
>am surprised that people got any different ideas from reading the
>document or looking at Figure 1. The resources are managed by the
>hypervisor so this is the natural authority for the information
>exposed by the MIB module.
>
>Given this, I do not really see why the security template text is not
>sufficient.  If you expose the content of the MIB widely, then it does
>not matter whether it is read from a VM on the hypervisor, a VM on a
>different hypervisor or any other system with access to the Internet.
>(Sure, if something on a VM wants to break out of its walled garden,
>it might in principle help to know the hypervisor technology etc. But
>frankly, it is far easier and more robust to simply try all different
>break out techniques known to an attacker or to infer likely
>technology via fingerprinting or even social and web engineering).
>
>My preference is to make no change to the text. But, if the IESG feels
>strongly, I won't object putting in additional (but in my view
>unneeded) words if that helps to avoid spending cycles on discussions
>that I find not really relevant. (Who cares whether the MIB module is
>implemented on the hypervisor as part of a monolithic agent or a
>subagent - the security trade offs are known and documented - at least
>for the standardized protocols).
>
>/js
>
>-- 
>Juergen Schoenwaelder           Jacobs University Bremen gGmbH
>Phone: +49 421 200 3587         Campus Ring 1 | 28759 Bremen | Germany
>Fax:   +49 421 200 3103         <http://www.jacobs-university.de/>
>
>_______________________________________________
>OPSAWG mailing list
>[email protected]
>https://www.ietf.org/mailman/listinfo/opsawg

Attachment: smime.p7s
Description: S/MIME cryptographic signature

_______________________________________________
OPSAWG mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/opsawg

Reply via email to