[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: draft of the ISAKMP/Oakley draft
Hilarie,
> I think you've got the patent situation backwards. It's GF[2^n] that
> is unencumbered.
I know that Certicom has patents in the field of
elliptic curves over GF(2^N). I don't know patents for
elliptic curves over GF(p). If you know some, please tell me.
> The available performance information indicates GF[2^n] can be exceedingly
> fast; I've not seen anything documenting comparable performance gains
> for GF[p].
>
> The GF[2^N] arithmetic is not a killer ... who can object to shift and XOR?
>
I do not object to shift and XOR, but an implementation
of elliptic curves over GF(p) does not require any additional
field arithmetic because the mod p arithmetic has always to
be implemented.
I do not suggest to remove GF(2^N) curves I suggest to add GF(p)
curves.
GF(2^N) arithmetic is very fast in hardware.
Perhaps you know running times showing that software realizations
of curves over GF(2^N) are also considerably faster than those of
curves over GF(p). If so I am very interested in these numbers.
Michael
Follow-Ups:
References: