Avaliação do vCubeChain em Sistemas Sujeitos a Falha Crash

Abstract

Blockchain is a secure and constantly growing system of transaction records in which each data user maintains a copy of the records that can only be updated if all parties involved in a transaction agree to the update. That is, it is a distributed peer-to-peer ledger that is cryptographically secure, immutable, and updatable only by consensus or peer agreement. Several consensus alternatives have already been proposed. The most recent work in this area is vCubeChain, a scalable permissioned blockchain solution. Based on vCube’s virtual topology - a distributed diagnostic algorithm that virtually connects the nodes of a network - vCubeChain uses the fault detector provided by vCube, which forms a hypercube when all processes are correct, to maintain logarithmic properties even when processes fail. In the article about vCubeChain, a series of experiments were presented to showcase the performance of the solution and compare it to well-known blockchains such as Ethereum and Bitcoin. However, the tool used does not allow the simulation of scenarios in which the network nodes fail. Therefore, in order to analyze the performance of the vCubeChain in the presence of failures, this paper proposes additions to the BlockSim simulator that extend it to simulate process failures. Furthermore, the HyperLedger Fabric, a permissioned blockchain strategy, is implemented to compare and analyze scenarios between public and permissioned blockchains. The results show that vCubeChain’s number of exchanged messages is much smaller than its counterparts, even when the leader node fails and a new leader election happens - leading to a larger number of messages being exchanged between the nodes to propagate this information.