Group structures of elliptic curves over finite fields

Abstract: It is well-known that if $E$ is an elliptic curve over the finite field $\mathbb{F}_p$, then $E(\mathbb{F}_p)\simeq\mathbb{Z}/m\mathbb{Z}\times\mathbb{Z}/mk\mathbb{Z}$ for some positive integers $m, k$. Let $S(M,K)$ denote the set of pairs $(m,k)$ with $m\le M$ and $k\le K$ such that there exists an elliptic curve over some prime finite field whose group of points is isomorphic to $\mathbb{Z}/m\mathbb{Z}\times\mathbb{Z}/mk\mathbb{Z}$. Banks, Pappalardi and Shparlinski recently conjectured that if $K\le (\log M)^{2-\epsilon}$, then a density zero proportion of the groups in question actually arise as the group of points on some elliptic curve over some prime finite field. On the other hand, if $K\ge (\log M)^{2+\epsilon}$, they conjectured that a density one proportion of the groups in question arise as the group of points on some elliptic curve over some prime finite field. We prove that the first part of their conjecture holds in the full range $K\le (\log M)^{2-\epsilon}$, and we prove that the second part of their conjecture holds in the limited range $K\ge M^{4+\epsilon}$. In the wider range $K\ge M^2$, we show that a positive density of the groups in question actually occur.