The Design And Implementation Of Performance Attribution Platform Based On Microservice

After more than 20 years of development,the fund has become a well-known and gradually accepted investment tool,which satisfies the wealth management needs of the public and institutional investors such as banks.With the increase in the number and types of fund products,when investing and financing,how to reasonably evaluate the performance of a fund,how to evaluate fund managers' real investment ability and the management ability of fund products,are the focus of mutual concern of fund companies and investors.The performance attribution analysis can provide a scientific explanation for these problems.Domestic theoretical research on performance attribution started late,although the third-party institutions that study fund performance have emerged one after another,but because of the lack of authenticity and coverage of data acquisition,they are only in the research and experimental stage,and it is difficult to provide personalized services for fund companies.The traditional performance attribution system of early outsourcing is also difficult to support rapid and frequent changes in business needs due to old technology.Based on the above background and the actual development needs of the com-pany,this thesis designs and implements a self-used performance attribution platform,which is divided into core modules such as performance management,attribution anal-ysis,transaction daily report,background management and other modules according to functional requirements.The platform is developed on the basis of Spring Cloud tech-nology with microservice style.It builds microservices around the business,which increases the flexibility of the system and ensures the scalability of the system under the trend of business innovation and technological innovation.Front-end interface is mainly developed based on Vue.js framework.The MVVM mode solves the problem that the business logic code and the view code are mixed and difficute to maintain in traditional development mode.The platform combines Java,the mainstream system development language,with R,a powerful statistical analysis tool,and uses ECharts,a powerful open source library,to realize the visual display of mass data,to make up for the deficiencies of the traditional performance attribution platform in data analy-sis capabilities and visualization.In addition,the thesis studies the way of calling R resources through MLServer(Microsoft Machine Learning Server),which solves the blocking problem of single thread.At present,the microservice-based performance attribution platform studied in this thesis has been officially put into use,and with its excellent features such as high availability,practicability and ease of operation,it provides good support for platform users in their business work.