John


I don't believe this iis  an off-line subject. As I recall we agreed to defer 
discussion of "Bypass" to an off-line topic. The topic we are discussing 
concerns what constitutes high assurance. I am simply pointing out that the 
"COMPUSEC" policy you were referring to was an access control policy and that 
the specifics of a given policy have nothing to do with the definition of what 
is or isn't high assurance.  Assurance has to do with the strength or 
rubustness of the measures used to enforce any given set of security policies.  
As a practical matter, the security policy does have bearing because unless the 
assets being protected are considered valuable (either intrinsically or 
otherwise) it is unlikely that the related security policies will be very 
strong nor will it be considered as cost-effective to invest in measures that 
would provide high assurance.



Another aspect of high assurance is that whatever is being certified/evaluated 
will almost certainly involve one or more independent evaluation groups as part 
of the process.



I'd be interested in other's views on this topic.



Thanks



John F.



________________________________

From: Davidson, John A. [[email protected]]
Sent: Tuesday, May 24, 2011 1:44 PM
To: [email protected]
Subject: Re: [cicm] BoF Request for CICM at IETF 81


John, That involves that thing we agree to disagree on.  Lets please take that 
continuing discussion off line.  Thanks, John


----- Original Message -----
From: [email protected] <[email protected]>
To: CICM Discussion List <[email protected]>
Sent: Mon May 23 17:52:45 2011
Subject: Re: [cicm] BoF Request for CICM at IETF 81

John, I don't believe that was quite true. I believe you are interpreting 
Bell-LaPadua as an overall security policy, when in fact is was strictly an 
access control policy as were most COMPUSEC policies.  The US Government's 
classification system is in fact part of an access control policy.  High 
Assurance is achieved when it can be shown that the security policy for a 
platform is rigorously enforced by the security mechanisms...regardless what 
that security policy may be.  Even the access control policy which says "no 
flow down" can be "bent" in real world situations when the situation demands.  
Things are never just black and white (or perhaps I should say black and red).

________________________________
From: Davidson, John A. [[email protected]]
Sent: Monday, May 23, 2011 10:06 AM
To: [email protected]
Subject: Re: [cicm] BoF Request for CICM at IETF 81


The conventional COMPUSEC view of high assurance was that - it was indicated 
where the Policy had to be enforced for certain (mandatory) e.g. no flow down 
tolerated.

----- Original Message -----
From: [email protected] <[email protected]>
To: CICM Discussion List <[email protected]>
Sent: Mon May 23 05:27:45 2011
Subject: Re: [cicm] BoF Request for CICM at IETF 81

Richard,

On 2011-05-22 at 06:36, Richard Graveman wrote:
> It seems to me that high assurance may well be needed in cases with
> only one domain. Is that out of scope?

Single domain use cases are definitely in scope; but they are very similar
(conceptually) to existing commercial crypto APIs. The ability to separate
domains is what sets CICM apart.

See:
"2.3. Single Security Domain" in CICM Logical Model
http://tools.ietf.org/html/draft-lanz-cicm-lm-00#section-2.3

"18. Single-Domain" in CICM Channel Management
http://tools.ietf.org/html/draft-lanz-cicm-cm-00#section-18

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

Reply via email to