HBSBA: Design of a Hybrid Bio-Swarm model for enhancing Blockchain miner performance through resource Augmentation techniques

Abstract

Blockchain mining is a power &resource consuming task, which requires multiple-levels of optimization, both at resource &task level. Over the years, a wide variety of mining optimization models are proposed by researchers, but most of them are applicable only to a subset of mining types. For instance, mining models used for Proof-of-Work (PoW) consensus-based mining, are not applicable for Delegated Proof-of-Stake (DPoS), and other consensus types. This limits the scalability of these models, which reduces their adoptability for dynamic blockchain systems (DBSes). These DBSes utilize different consensus models as per context of data storage, and are widely used by blockchain designers to deploy high-efficiency, and low delay storage solutions. A standard mining optimization solution is not available for such scenarios, due to which researchers & system designers opt for deployment-specific optimizations, which need to be redesigned for each blockchain system. To remove this drawback, a standard blockchain mining optimization model is proposed in this text. This model uses a combination of Genetic Algorithm (GA) & Particle Swarm Optimization (PSO) for solving two different issues. The GA model is used to optimize miner set selection, which will be used for consensus, while the PSO model optimizes the responses from these miner sets depending upon their temporal mining performance. Due to optimum miner set selection, only higher efficiency miner nodes are used for mining the blockchain. While due to performance optimization of these miner nodes, their internal mining efficiency is improved.This efficiency is evaluated in terms of delay & power needed for single block mining w.r.t. blockchain length. It was observed that a combination of these models is capable of enhancing mining speed, with reduced power consumption, and higher mining throughput. Due to this improvement the proposed HBSBA model outperforms most of the recently proposed blockchain mining models. The model was evaluated on DPoS, Proof-of-Authority (PoA), Proof-of-Stake (PoS), and PoW based consensus models, and a delay reduction of 14.5%, throughput improvement of 8.3%, and reduction in energy consumption by 4.6% when compared with various state-of-the-art models. Due to this improvement, the proposed model is applicable for a wide variety of medium to large scaled blockchain mining applications.