|
|
I-0389: Recovery To A Known State |
TYPE: NIAP Interpretation
NUMBER: I-0389
STATUS: Formally Rescinded
REASON: CCIMB Rejected Interp. See February 2003 NIB Agenda Item
1.d.iv.
TITLE: Recovery To A Known State
EFFECTIVE: 2001-03-15
RESCINDED: 2003-05-09
SOURCE REFERENCE: CC v2.1 Part 2 Subclause 10.8 FPT_RCV
CC v2.1 Part 2 Subclause J.8 FPT_RCV
RELATED TO:
I-0406 Automated Or Manual Recovery Is Acceptable
CCIMB ENTRY: CCIMB-INTERP-0149
ISSUE:There are situations where some form of recovery from a known backup is required, but there is no formal model to argue that the known state is provably secure.STATEMENTIt must be possible to recover to a known previous state, as opposed to one that is provably secure.RECOMMENDED CRITERIA CHANGESTo address this interpretation, the following changes are made to CC v2.1,
Part 2:
(additions marked
thusly; deletions marked
SUPPORT:The words in the annex for FPT_RCV state:Throughout this family, the phrase "secure state" is used. This refers to some state in which the TOE has consistent TSF data and a TSF that can correctly enforce the policy. This state may be the initial "boot" of a clean system, or it might be some checkpointed state. The "secure state" is defined in the TSP model. If the developer provided a clear definition of the secure state and the reason why it should be considered secure, the dependency from FPT_FLS.1 to ADV_SPM.1 can be argued away. Although this allows a secure state to be a previously checkpointed state, this ability is buried. This component makes it explicit; it also makes it clear that a previously known state may or may not be a secure state. Note: This interpretation is being applied to the CC as modified by I-0406. 2003-07: The CCIMB issued the following statement regarding this interpretation: The national interpretation is an incorrect interpretation of the CC: The RI problem statement assumes that one needs a formal Security Policy Model (ADV_SPM.3) to define secure states. This is an incorrect assumption: an informal Security Policy Model (ADV_SPM.1) is sufficient to define the secure states of a TOE. The current components in FPT_RCV all have dependencies on ADV_SPM.1, implementing this fact. CC Part 2 para 1236 expands on this by stating that if one can give a rationale why a particular state is secure, a full Security Policy Model is not necessary (and hence the dependency on ADV_SPM.1 can be deleted). The suggested FPT_RCV.NIAP-0389-1 Recovery to Known State allows return to a previously known state, which cannot be considered to be a secure state (otherwise the author would use the stronger FPT_RCV.1). This means that the inclusion of FPT_RCV. NIAP-0389-1 may create vulnerabilities in a TOE, as attackers may use the fact that the TOE returns to a previously known state that they have reason to assume is insecure, as the following example illustrates: "A TOE stores its state every five minutes and, in the case of failure/service discontinuity, restores its previous state. If an attacker induces a failure in the TOE within five minutes of an authorised user logging, the TOE restores the last stored state, and now thinks that the authorised user is still logged in." |