Bill Sommerfeld wrote: > On Thu, 2007-08-23 at 07:47 -0600, Mark A. Carlson wrote: > > I don't think we should be making examples of cases, > > nor forcing each one to invent their own approach. > > If that's the case then we should just pack up and go home, because all > a project team needs to get approval from us is to assert ETOOHARD > enough times to wear us down. > > The first user of a new facility will be copied, so it should get the > details right. > > > I believe we can say that the read protection provided by 2007/177 > > meets the spirit of the policy until we change or abolish the policy > > itself. > > It's not our policy, so "we" (PSARC) cannot unilaterally change or > abolish it. > > > Lame reversible obfuscation sounds like "security through obscurity" > > to me. > > The policy specifically calls for obfuscation as an alternative when > stronger measures aren't possible. It's not for us to unilaterally > interpret the policy to delete that provision.
If a simple obfuscation would suffice, that's not a problem and we agreed to do that some time ago. A separate project to provide a common, key-based security infra- structure for 2007/177 makes a lot me sense. My concern has been the value of implementing a key management infrastructure specifically for the NDMP project. If we obfuscate the NDMP password using base64 encoding within a randomized, fixed-size, 256-byte property buffer, would that be sufficient? Alan
