Performance And Security Analysis Of Blockchain For Internet Of Things

Blockchain has been regarded as a promising technology for Internet of Things(Io T),since it provides significant solutions for decentralized networks that can improve the security of network,alleviate the pressure of high-load nodes and achieve adaptive matching and behavioral decision-making of Io T device.Considering the Io T context usually has the characteristics of limited resources,large scale,and high network load,the consensus algorithms used in Io T should meet the needs of high security,high resource efficiency,high scalability,low cost,and need to provide sufficient transaction throughput.As a result,how to find a consensus algorithm that matches the Io T context has become a hot and meaningful topic.In this work,based on the needs of the Io T context,we discuss the main ideas of famous consensus algorithms,including Proof of Work(Po W),Proof of Stake(Po S),Practical Byzantine Fault Tolerant(PBFT),Directed Acyclic Graph(DAG)-based consensus algorithm,and list their advantages and limitations in Io T.Then,we propose mathematical models to study the performance and security of consensus algorithms.The main content and contribution of this work are listed as follows:Firest,considering the DAG-based consensus algorithm(referred to DAG consensus)has the advantages of high transaction throughput,low resource consumption,and zero transaction fees in Io T context,based on one of the most typical DAG consensuses,Tangle,we propose a Markov chain model to investigate the impact of Io T network load change on the system performance.For the securiy,we analyse the attack strategies that a rational attacker would take under different load regimes.Then,we leverage a stochastic model to analyse the probability of a successful double-spending attack.The results can provide an insightful understanding of DAG consensus process,e.g.,the network load has significant impact on the system performance and secutiy,and a lower cumulative weight growth rate will result in a higher successful probability of double-spending attack.Meanwhile,we also demonstrate the trade-off between security level and confirmation delay,which can act as a guidance for practical deployment of blockchain systems.Then,we focus on the transaction throughtput limitation and high recource consumption in Po W algorithm,and propose a Block Access Control(BAC)approachto address the forking problem in wireless Io T context,which aims to achieve high transaction throughtput as well as block utilization in Po W consensus process.Using Markov chain model,we evaluate the performance of the BAC-based Po W algorithm,including transaction throughtput,block discard rate,block utilization,and mining pause probability.The results show that the proposed approach enable a high transaction throughtput for Po W algorithm in wireless Io T context.Meanwhile,the block size and Po W hash difficulty will influence the system performance and security.This analysis provides a guidance to configure the blockchain system which can achieve a trade-off between transaction throughput and block utilization.