Splitting and Aggregating Signatures in Cryptocurrency Protocols

The blockchain technology and a vast amount of cryptocurrency related activities have generated an unprecedented level of interest among the public. However, even at the entry level, cryptocurrency users need to deal with the complex task of key management. In this paper, we propose a simple way to manage a user's private key, under a reasonable assumption that the user has two devices at his disposal (say a laptop and a mobile phone). We refer to our strategy as key splitting. Since these cryptographic keys are used for generating digital signatures, we should take a closer look at the signature schemes that would perform best under key splitting. At the operational level, scalability is one of the main challenges faced by the users and developers. While there are fundamental issues like consensus that challenge scalability, we focus on the computational efficiency in a block formation. Aggregation of signatures is one of the effective solutions to this problem. To this end, we observe that none of the existing signature schemes work well for BOTH key splitting and aggregation. The current popular schemes such as the ones used in Bitcoin or Schnorr's scheme implemented over Elliptic curves are neither suitable for aggregation nor can their keys be split in a convenient and meaningful way. A detailed theoretical and empirical analysis shows that the BLS short signature scheme is best suited for achieving both key splitting and aggregation.