The prime factorisation problem is intractable, i.e., no efficient algorithm is known. In cryptography there are some well-known approaches based on the computational hardness of this problem, including Rivest-Shamir-Adleman (RSA) encryption protocol. Several attemps to break RSA have been investigated, some of them based on the massive parallelism of membrane systems.In this paper a new approach, based on kernel P system formalism, aimed at reassessing the space-time tradeoff, usually involved in membrane computing solutions, is investigated. Two models are introduced and assessed in order to find the potential benefits of each of them for solving the above problem, revealing also their limitations and providing hints for further improvements.