| Blockchain is a distributed system featured with traceability of information,immutability of data and transferability of value.Thanks to the rapid development of blockchain technology,recent years have seen the mushroom growth of new business models such as digital currency transactions and trusted resource chaining,and large quantities of blockchain-related data have thereby been generated.As a new type of data resources,they have drawn attention from both academic community and industrial world.A major research focus at present is applying various technologies to the analysis of blockchain data,among which are automatic analysis by intelligent algorithms and visual analysis by human-computer interaction,providing a great impetus to the deeper understanding of the core of blockchain technology and the expansion of its application.This dissertation addresses blockchain-related data from the perspectives of the intelligent analysis of data,the visualization of data,and the integration of blockchain and visualization technology,focusing on such main scenarios as digital currency transactions and trusted resource chaining.The main components of this research are as follows:1.Research on the detection method of money laundering behavior in digital currency exchange.Digital currency transactions based on blockchain technology mostly take place in digital currency exchanges,which support not only coin-currency trading between digital currencies,but also the exchange of digital currencies with fiat currencies,in a low-cost,fast,highly anonymous,and decentralized manner,lending themselves to becoming new places for money laundering.The existing anti-money laundering techniques are designed for traditional financial practices,difficult to handle the emerging financial form of digital currencies.This dissertation puts forward an approach to detecting money laundering behaviors in digital currency transactions.It first defines the digital currency transaction behaviors,develops a framework to describe their features,designs an improved local anomaly factor algorithm to detect,before quantifying the degree of,abnormal transaction behaviors,and finally identifies the suspected money laundering users by means of the results thereof.This dissertation conducts multi-algorithm comparative experiments on the real transaction data(3 million transaction records of 10,000 users)in two years in an S Digital Currency Exchange.The results demonstrate that the method is effective in detecting significant and hidden abnormal transaction behaviors in the Exchange in question,helping its security administrators spot highly suspicious money launderers and improving their efficiency of anti-money laundering investigation and evidence collection.2.Research on the visualization method of blockchain transaction data.Digital currency transactions based on blockchain technology are stored with a blockchain-specific bookkeeping method.Most of the digital currency retrieval systems in use present only common texts and lists for users to view transaction data,which is limited in expressiveness,not friendly to beginners,and not informative enough for experienced users.This dissertation designs a novel visual browser of blockchain transaction data by introducing such visual metaphors as paper ledgers and copper coins,and adopting four visual designs for four data objects: blockchain,block,transaction and address.The browser represents a mode of lightweight interactive data exploration,facilitating beginners’ easy understanding of blockchain-related concepts and quick retrieving of transaction data,and providing experienced users with an effective access to more advanced transaction information.The visual browser was successfully applied to the Silubium digital currency search system from May 2018 to May 2020,with more than 30,000 users using it for browsing during the peak period.We have now open-sourced the source code of this browser(see list of results)to promote the application and spread of visualization technology in the field of digital currencies.3.Research on the application of blockchain and visualization technology in campus network management.To address the difficulties in locating end-users and solving technical failures in the routine management of university campus network,while helping CTBU upgrade its campus network management platform,This dissertation contributes a user-centered multi-space network security visual analysis method and a synergistic visual analysis scheme of network geo-space,network topological space and network IP space,with a view to opening up channels for information to flow among multiple network spaces.This dissertation also designs a network management information chaining method based on blockchain and smart contract technology to store the configuration information of core campus network devices into the blockchain,so as to achieve trusted chaining and secure traceability of core information resources and prevent the core network configuration of campus network from being tampered with unintentionally or maliciously.The application research focuses on the human-centered idea by involving several actual users in the whole process of demand analysis,solution design,system development and application evaluation.Positive feedback is heard from them,and their efficiency of routine campus network management is improved substantively.In addition,This dissertation investigates the optimization methods of blockchain consensus algorithm.The blockchain consensus algorithm directly affects the performance,security and stability of the blockchain system.This dissertation proposes an optimization method of proof-of-stake consensus algorithm to deal with the problem of unstable block-out time,specifically,using the elected master node and the defined fastest and slowest block interval time,and then monitoring the block packing process.If the proof-of-stake consensus algorithm does not generate blocks in the expected time,the master node will generate blocks directly.This optimization method can enormously improve the performance of the blockchain system based on the proof-of-stake consensus algorithm,so that each transaction can be packed and confirmed in the fastest manner,with much higher stability of the blockchain block-out.Figure:24,Table:18,References:137... |
PDF Full Text Request