Blockchain Data Collection And Analysis Framework

Blockchain has attracted widespread attention since its invention.Many features such as tamper resistance,traceability,anonymity and openness have attracted more and more people's interest.Many researchers and practitioners in related industries have set foot in this field.If people want a deeper understanding and more detailed analysis of the blockchain,they need to complete data collection.As of now,the two most well-known blockchain platforms are Bitcoin and Ethereum.This thesis will implement a blockchain data collection and analysis framework based on Ethereum.Ethereum is the largest blockchain platform supporting smart contracts with the sec-ond biggest market capitalization.Exploring it can yield many useful insights because of the huge volume of transactions,accounts and blocks as well as the popular applications developed as smart contracts,not to mention many new attacks on the platform and the smart contracts.Unfortunately,it is non-trivial to systematically explore Ethereum be-cause it involves massive heterogeneous data,which are produced and stored in different ways.Although a few recent studies report some interesting observations about Ethereum,they are limited by their data acquisition methods which cannot provide comprehensive and precise data.In this thesis,to fill the gap,we propose a systematic and high-fidelity data exploration framework for Ethereum by exploiting its internal mechanisms.Besides supporting the analysis in existing studies,further empowers users to explore unknown phenomena and get in-depth understandings.In order to demonstrate the practicability and effectiveness of the data collection function,this thesis implements multiple applications from three dimensions including Ethereum entities,smart contracts,and tokens.They have conducted in-depth analysis of Ethereum from multiple perspectives.Applications include account balance analysis,smart contract bytecode statistics and call library analysis,DAO reentry attack detection,non-deployable contract detection,underpricing DoS attack detection,and token inconsis-tent behavior detection.Based on the experimental results,we summarized the behavioral characteristics of the mining pool,the characteristics of the contract bytecode and the call library? discovered the unreported DAO contract attack surface? analyzed the use of non-deployable contracts? discovered unreported DoS attack transactions and found multiple non-standard tokens and their inconsistencies.The thesis first describes how we solve the challenging problems in the data collection module,then introduces the application and experimental analysis of the data analysis module in detail.The token behavior detection and analysis is a complex application,which will be introduced in a separate chapter.At the end of the thesis,this work is summarized and the future content is expected.