Re: A very fast Fermat factoring algorithm

From: none (_at_(none))
Date: 04/04/05 

Date: Mon, 04 Apr 2005 13:07:52 +1200

Erwann ABALEA wrote:
> Bonsoir,
>
> On Fri, 1 Apr 2005, it was written:
>
>
>>Erwann ABALEA wrote:
>>
>>>On Thu, 31 Mar 2005, it was written:
>>>
>>>
>>>>I use 4k bit keys. I let 1k bit keys expire rather quickly..
>>>>
>>>>Computers are fast and cheap enough nowdays. There really is no reason
>>>>not to use them.
>>>>
>>>>I would like to think that CA root keys are in the 8K+ bits range.......
>
> --
>
>
>>>4k bits at most, but generally 2k bits max:
>>> - speed problem
>>
>>Speed problems? what speed problems.. OK on a HW token sure. But on
>>modern PC's. I don't think so.
>
>
> I hope you won't trust a CA that operates its private keys in software on
> a PC... Everything I said is valid for a CA, not for an end-user. Speed is
> not an issue for an end-user, because generally the slowest part is the
> user.
>
Yes of course - I just didn't think it though far enough in regards to
the CA specifics. Key compromise is the number one threat without a doubt.

OTOH i don't really trust most of the current CA's. You just need to
read there terms and conditions. Or just the procedures to get a cert.
Not too hard to get Verisign to give a cert when your not who you are
suppose to be.

But case in point, with modern hardware there is little need to use
restrictive tokens...... Use very powerful one .. something more like a
PDA than a smart card.

greg

Loading