I am trying to implement a basic Diffie-Hellman key exchange using ECC and HECC of genus 2 and 3 and compare them. However, I am a bit confused about the concept of key size in HECC.
In ECC, from my understanding so far and considering I am new to the field, the private key size is basically the bit size of the prime that is used to define the prime finite field, where the elliptic curve is defined.
In HECC, instead of a point in the Curve, a Divisor class is used to define group elements. In my implementation, I have defined divisors in the Mumford representation, which is basically 2 polynomials u, v.
Based on this paper: https://www.researchgate.net/publication/305878169_A_Study_of_Hyperelliptic_Curves_in_Cryptography
one can see that the key size drops while the genus g increases. However, I cannot actually understand what we define as a key size in HECC. Logically, the coefficients of the polynomials u, v define the private key, so the sum of their bit sizes should give us the key size. However, even if we choose a smaller finite field in HECC, because the degree of the polynomials u, v increases as the genus increases, that should not mean that the key size increases as well?
So if I want to compare ECC, HECC genus 2 and HECC genus 3 with a security level of 2048 for example, what is the bit size of the prime finite field that I should choose and what will be the actual key size in each implementation?