Research On Data Provenance Based On Blockchain

Bitcoin as the first application based on Blockchain was invented by Satoshi Nakamoto in 2009.The conception of Blockchain and those key technologies such as decentralization,transparency and reliance spread all over the world increasingly in 2015.Blockchain is an Internet-based distributed ledger technology.It is a decentralized or multi-centered database and a huge chain consisted of numerous data blocks linked one by one.According to the process on generating a block,Data could be classified into two categories,current data which should be accepted by all nodes in distributed system and historical data stored in block chain permanently.This process is to turn current data into historical data.As depicted above,the two key technologies about Blockchain are consensus and chain struct.This essay is to research consensus and chain struct for enhancing the efficiency on data tracing.An improved consensus algorithm based on Practical Byzantine Fault Tolerance(PBFT)and a new chain struct are proposed in this paper.The main objectives of this essay are described as follow:Group-Hierarchy(GH)as the improved PBFT focusing on reducing message complexity is designed to advance efficiency on completing one request from client.Analyzing the message complexity of GH and PBFT theoretically and GH has less message complexity than PBFT.GH is the basis of the distributed data trace on communication in this essay.Simulating a large distributed system with 100 nodes to test the consensus time about GH and PBFT and the result agrees with the estimationDividing traditional one chain into three chains about Users' states,Merchandises' states and Transactions' state and each chain is associated with each other.At the same time,data in the traditional chain is divided into two parts,one is block header stored in file and the other one is block body stored in database.Block header is used to verify and block body is used to trace specific data.The three chains is the basis of the distributed data trace on storage in this essay.Setting up a small distributed system for testing those combinations,GH-One chain,GH-Three chains,PBFT-One chain,PBFT-Three chains.The result shows that GH-Three chains has the least consensus time.