The Desigh And Implementation Of Real-time Domestic Market Data Platform

With the development of electronic technique and the change of investors' trading customs,program trading has gradually become an important direction of the development of the securities market in China.In the context of increasing trading frequency,the sooner to acquire market data,the more likely to succeed in trading and get opportunities to generate huge profits.Therefore,the convenience of acquiring the whole market data and the low latency of the data are particularly important.Thus,it is an important research direction to design and implement an efficient and practicable high-speed market data platform.Firstly,this thesis makes a systematic analysis on the high-speed market data platform,clarifies system requirements and objectives and expounds the feasibility of the implementation of the system.Secondly,in order to achieve the function requirements and performance requirements of the system,this thesis analyzes the key technologies,including the exchange market data access and analysis,asynchronous network communication,multithread processing and memory pool technology.Based on the previous analysis and summary,this thesis determines the design and implementation of the market data platform.Aiming at the problem of the imperfection of existing market data,this system selects the Level2 market data in Shanghai and Shenzhen Stock Exchange and the Femas market data in China Financial Futures Exchange as high-quality data resources.Using parallel analytical methods,this system covers all the market data of all exchange-traded products.In terms of the latency of the market data,this thesis puts forward the solution of rapid extraction of STEP messages and FAST code data based on parallel analysis,significantly speeding up the parsing speed for market data.In order to improve the efficiency of using the market data,the system support the LAN-based multicast of market data based on the subscribe manners,which significantly reduces the pressure on network bandwidth under multi-user scenarios.Meanwhile,this system provides a unified market data interface to support multiple programming languages for investors,highly reducing the cost of adaptation for users.Finally,in order to verify the efficiency of the whole variety market data platform proposed by this thesis,the system is tested from the point of view of function and performance.In function test,this thesis mainly verifies the accuracy of the market data and the automatic diagnosis ability of system single point failure.In performance test,by comparing this platform with other platforms in terms of market data delay,this thesis verifies the low market delay and better data stability of this platform.So far,the feasibility and effectiveness of the method are verified by test and actual operation.