Design And Implementation Of Program Futures Trading System

With the rapid development of Chinese economy,the futures market has undergone enormous changes.Manual trading mode can not meet the development trend of futures trading increasingly,so the program trading emerged.Although the program trading started late in our country,more and more people realize the importance of program trading.Program trading has high requirements in terms of real time,but also requires a high degree of flexible strategy.If you take into account factors such as risk monitoring,design complexity in program trading systems will be greatly enhanced.So current problems to be faced in the implementation of program trading systems is how to design and develop a program trading system which not only to meet the high degree of flexibility in the trading,risk control,strategy,but also to ensure a high degree of stability.Closely based on the futures trading characteristics and combined with the existing shortcomings in terms of program trading,the paper proposed the implementation of a complete set of solutions futures program trading system,which has a flexible and easily customized back-tested features.Program trading system is divided into four major subsystems,including trading subsystem,risk control subsystem,policy subsystem,database subsystem.Trading subsystem consists of three major components,including transaction initialization module,data exchange module,core transaction module;wind control subsystem consists of risk control management monitoring layer,risk control strategy metering layer,database layer,and data layer;policy subsystem consists of policy control module and policy running module;database subsystem consists of Oracle database,Mongodb databases,in-memory database.Each part of the design and implementation instructions is given in the paper.In this paper,in-depth study of high-speed trading and risk monitoring.The transaction subsystem uses a modular design,and in the aspect of high-speed transactions,according to futures trading rules,the design ensures that the system has high throughput,low latency characteristics;To ensure the correct state to switch,internal communication in subsystem uses state machine design pattern which ensures that the system has high availability and stability.In terms of risk monitoring,the paper introduced non-relational in-memory database Mongodb based on big data processing and Complex Event Processing(CEP)technology.Mongodb is mainly used to solve the problem of storaging unstructured data;CEP is to solve the real-time risk control requirements.By combining them,a better regulatory system solution of the design problem for a system including multi-product,multi-contract,cross market and sub-account.Finally,the classic R-Breaker policy is achieved on the program trading system,Test results,which achieved the largest overall gain 188% and maximum retracement of 7%,using 2010-2012 data,validate the program trading system in terms of business logic;While stress tests in terms of performance trading systems has been taken.Test results show that in the case of 4000 users simultaneously log in and 5000 orders per second,program trading system still has excellent usability and stability.The overall test results show that program trading system design goals have been achieved.The design and implement of the system have now been completed.The system performs better in the test environment,meeting the business objectives and enhances the futures company's risk monitoring capabilities.