Score:1

What is the order of the generator point G=9 in curve25519?

vn flag

In Curve25519 we typically have this generator point or base point:

Gx = 9
Gy = 14781619447589544791020593568409986887264606134616475288964881837755586237401
or:
Gy' = p - Gy 
   = 43114425171068552920764898935933967039370386198203806730763910166200978582548

Where p = 2^255-19, the dimension of the prime field Fp in which we evaluate the curve.

What is the order of this generator point?

i.e. what is the smallest n so nG = 0.

Before actually thinking about it, I just assumed that would be p since p is prime. But obviously that's wrong as we're dealing with elliptic curve point addition here, not just scalar multiplication in modular arithmic.

So I'm wondering what is G's order, and perhaps more difficult: how can I find this myself? (once I have the value I can easily verify it, that's much less complicated)

knaccc avatar
es flag
$2^{255}-19$ is not the curve order, it's the dimension of the prime field. The curve order is the number of possible points on the curve, which is $8p'$ where $p'=2^{252}+27742317777372353535851937790883648493$
RocketNuts avatar
vn flag
Thanks for correcting my careless misnomer, yes of course the p in Fp (or Z/pZ) has nothing to do with the curve.
kelalaka avatar
in flag
[Summarize the mathematical problem at the heart of breaking a Curve25519 public key](https://crypto.stackexchange.com/a/50414/18298)
kelalaka avatar
in flag
Dupe for finding the order [How is the order of a point calculated for elliptic curves over GF(p)](https://crypto.stackexchange.com/q/40726/18298) and above was the dupe for the title.
Score:5
cn flag

According to this source, the points of this curve are a group of cardinality $8\cdot p'$ with $p':=2^{252}+27742317777372353535851937790883648493$.

This number can be computed by using the Schoof algorithm or the more efficient Schoof–Elkies–Atkin algorithm.

Then, by Lagrange theorem, and because $p'$ is prime (can be check with any efficient Primality test), it implies all the points $P$ can only have order $o_P= 2^{i_P}\cdot p^{\prime j_P}$, with $0\leq i_P\leq 3$ , and $0\leq j_P\leq1$.

We can compute $p'\cdot G$ with the fast exponentiation (Square and multiply algorithm called also Double-and-Add in a elliptic curve context), and notice it's equal to $\mathcal{O}$ the neutral element of the curve.

We deduce that $o_G$ the order of $G$ divides $p'$. Then $i_G= 0$.

Because $G\neq \mathcal{O}$, $o_G\neq 1$, then $j_G=1$.

We conclude that $G$ is of order $o_G = p'$.

mangohost

Post an answer

Most people don’t grasp that asking a lot of questions unlocks learning and improves interpersonal bonding. In Alison’s studies, for example, though people could accurately recall how many questions had been asked in their conversations, they didn’t intuit the link between questions and liking. Across four studies, in which participants were engaged in conversations themselves or read transcripts of others’ conversations, people tended not to realize that question asking would influence—or had influenced—the level of amity between the conversationalists.