Re: MD5 irretrievably broken

2008-12-31 Thread Rob Stradling
On Tuesday 30 December 2008 22:07:08 Kyle Hamilton wrote: I would suggest requiring all new roots approved to state that they do not and will not use MD5 in any newly-minted certificate (except possibly in a configuration like the TLS pseudo-random function). FWIW, Comodo have never signed

MD5 irretrievably broken

2008-12-30 Thread Chris Hills
A presentation was given at this year's Chaos Communication Congress in which it was described how researchers were apparently able to produce authentic signed SSL certificates thanks to a handful of CAs who rely on MD5. If true, is it time to disable MD5 by default?

Re: MD5 irretrievably broken

2008-12-30 Thread Kyle Hamilton
I would suggest requiring all new roots approved to state that they do not and will not use MD5 in any newly-minted certificate (except possibly in a configuration like the TLS pseudo-random function). This is not yet policy, though it should be. (FWIW, this was known two years ago.) -Kyle H

Re: MD5 irretrievably broken

2008-12-30 Thread Florian Weimer
* Kyle Hamilton: I would suggest requiring all new roots approved to state that they do not and will not use MD5 in any newly-minted certificate (except possibly in a configuration like the TLS pseudo-random function). If they issue certificates for sub-CAs, they have no technical means to