Re: [OpenPGP] Key Transition
Grant Ingersoll <[email protected]>
| Newsgroups | gmane.comp.apache.community |
|---|---|
| Message-ID | <[email protected]> |
Another question: When updating my KEYS file (per http://www.apache.org/dev/key-transition.html#transition-export ), do I replace my old one with the new "dual" export, or do I append to the KEYS file? -Grant On Aug 19, 2009, at 6:07 AM, Robert Burrell Donkin wrote: > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA512 > > i plan to prepare a set of instructions for those who have a small key > and who want to transition to a longer one. most of this should be > uncontroversial. > > there is one area that i think needs some more consideration. i don't > think that there is a great rush to implement this so - unless anyone > jumps in with something i've missed - i'll get on with the rest whilst > this is discussed. > > a transition statement is a notice informing the world that the old > key > is being replaced by a new one. it is signed by both the new key and > the > old. for an example, see > http://www.jroller.com/robertburrelldonkin/entry/openpgp_transition_statement > . > > providing that the old key has not been compromised, a transition > statement allows those who trust the old key - and that it hasn't been > compromised - to resign the new key. > > for apache, the risk with recommending this mechanism is that it's > less > secure than signing after a F2F meeting - if a key is compromised then > transition statements could be published and keys signed in error. > however, without using transition statements, there is a risk that an > advance in cryptography will conclusively break SHA-1 or DSA before > the > new apache WOT is viable. > > if we decide to recommend transition statements then i recommend > asking > committers to broadcast transition statements using a independent > trusted communication channel which can be monitored by committers. > this > should provide more security than each person using an ad hoc > solution. > > for example, asking for statements to be committed to subversion would > mean that an attacker would have to comprise a users subversion > credentials as well as their private key. if that change were posted > to > a public mailing list to which that user were subscribed then they > would > be informed that their key had been comprised and could take > appropriate > action. > > opinions? > > - - robert > -----BEGIN PGP SIGNATURE----- > Version: GnuPG v2.0.11 (GNU/Linux) > Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org > > iQIcBAEBCgAGBQJKi87nAAoJEHl6NpRAqILLjqwQALzaot3LVWJwbCzxb2G9ZHbe > +1luE6cQ4BH8aEz/S8oZDYq0iekvmJESEYzylalis4H4NEMfoIvTKS5Wdthgwspj > IKxn6zjAgcj25+WFq+0sd8TK5BGoAYR9HOLkQsUEOFp3w693gbm3lE9XbRkFRMc5 > c/T9n4hVnPXGEih5fzaeHhOxGDcnuRGu4ZSs+GfW/F6hncqhTdqKw8kXTWeQ9es/ > 8xNcIkxULUIOHOgjVgEyQBHCX7zDsW7p3kBysHuYNV3BIKEwSOO660LmEUmnOLYR > PYqFMEMmpEL8BJYZvtz1b9CG/ROtBWmy7GsjiXAvClWvZw93w5O+/qwFZ6LYQgO2 > IRd+T+RknJzr7KdPE/vzrlCpAITNd5SU4ROpUT9hSj2cig7sZWwaPlC+W4fr+1eA > fk+PKPANEyBP2SnnNzmm9gOUCLahigHZVNR+8TBJVImAptQqvfpchrcwq+ov55vQ > AL/msg81DxZaj/TR3tjydy1xu61t2coJ1OAN/yn/UyeFxzyujHxdiHtudaCaAXeP > 7tfCvEvHa9q4DotvfT5aS6+hVQTUy2Hxd9iOHFwim6ewE2DVsvryeYI3PP60g/Vj > XLoE6vpkJn3TgObQrhnGzF9vKLRBptYFy0HK8BacOaVdP8oQGeX2/02AGJ7mEwTi > 7WpvJnmgD1ILnQt6ZrJx > =BVjD > -----END PGP SIGNATURE----- > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [email protected] > For additional commands, e-mail: [email protected] >