A Secure Blockchain-based Data Trading Ecosystem

Data,a key asset in data-driven economy and its increasingly important role has fueled the rapid development of the data trading industry.The data trading process relies on the data trading platform as the bridge between data buyer and data seller,the data seller's raw data flows to the data buyer with the help of data exchange.However,there are two important problems in the data exchange as following: First,A dishonest buyer may also resell the raw dataset after get raw data,without the original seller's knowledge and approval.Second,when the raw data of the data seller flows through the data exchange,data exchange may covertly cache and resell the seller's dataset,without the seller's knowledge and approval.In order to solve the security problems of the data trading platform,a Secure Blockchain-based Data Trading Ecosystem(SDTE)implements a security data trading mechanism.First,in SDTE,the data seller sends the raw data to the smart contract deployed by the data buyer on the Blockchain.SDTE also performs additional charging according to the size of the output data and disables calls between smart contracts to prevent the malicious smart contract from directly outputting the raw data.The buyer cannot obtain the raw data and ensures the security of the raw data.Second,raw data and execution results are encrypted,Intel Software Guard Extensions(SGX)is used to protect above operations and the running process of smart contracts.This not only ensures the security of the buyer and seller but also prevents the data exchange from acquiring their data and resells it.Not only that,in order to prevent the repudiation and fraud of roles in the system,SDTE deploys Data Trading Management Contract to constrain dishonest behaviors.To demonstrate the security of SDTE,the presence of security threats is modeled and detailed security analysis is performed.To demonstrate the practicability of SDTE,the data trading platform based on Ethereum and SGX is implemented to measure the additional performance overhead in the data trading process.The test result is based on the core module(using SGX to protect the smart contract execution)uses K-Nearest Neighbor(KNN)algorithm to process data,which shows the additional time is less than 0.5 s compared to the traditional smart contract execution process.