Based Middleware Adaptive Stock Trading Platform Design And Realization

In these days, the development of Chinese stock brokerage leads to extensive usage of stock concentrated trading mode, which is not only a significant measure of strategic adjustment to preserve competitiveness against counterparts and to revive the flagging stock market and shrinking brokerage revenue, but also a trump to success of all-round market competition in the future for securities brokers.A stock concentrated trading platform is a cluster of middlewares which is assembled by a number of trading middlewares. The cluster which serves the middlewares is responsible for topologic management, load balancing, failure monitoring and recovery and so on. It also provides high scalability, high reliability and high availability for the middleware platform.The trading platform should be a self-adaptive system which contains the feature of self-configuration, self-optimization, self-recovery and self-protection. This adaptive capability ensures the highly concurrent access in stock concentrated trading. It also improves the overall performance of the trading system.On the standpoint of improving adaptive capability for the stock trading platform, the paper explored the design and implementation of the adaptive stock concentrated trading platform, then designed and implemented adaptive topologic management module and adaptive load balancing module for trading middleware in business platform. As for adaptive load balancing module, dual-threshold load balancing algorithm is introduced and its shortcoming was also analyzed in details. And in order to solve the problems possibly caused by thrashing, the paper gave an improved solution for the algorithm.The testing results indicate that the adaptive trading platform can meet the requirements of highly bursting and concentrated access in stock business. By dynamically increasing the number of application servers, the system capability will be improved in linear growth. This can fairly relieve the symptom of performance bottleneck existed in trading system.