On Sat, 21 Apr 2007 10:48:05 +0200 Wouter wrote: WW> I have recently been thinking about key-priming for machines that have WW> been off-line long enough for their trust-anchors to go stale. [...] WW> * operators can discard old private key data and still do all of the WW> above. I've only skimmed the draft, but my first though when I read this was: what happens when an old key (which was never revoked) becomes compromised? The scenarios in section 6 of timers is 'informative', so a zone could roll keys without revoking the the keys as described in 6.3. So, does this present an attack on the stale update mechanism you describe? If so, then operators might need old private keys to go back and revoke the compromised key. Another thought I had was with regards to delegating the historical data to subzone. Why restrict it to a subzone? If you let it be any name, operators could outsource maintenance of the historical data. -- Robert Story SPARTA
Attachment:
signature.asc
Description: PGP signature