Construct Rich Client Bond Trading System

Recent several years, thought of Rich-Client is increasing. RIA(Rich Client Application), integrating with the response speed and rich functionality of desktop application and the wide usability of Web application, gives users great experience. Adobe Flex is one of the most mature products of rich client, and widely used as the client application of enterprise application. At the same time, in J2EE realm, as an light weight develop framework, Spring is growing very fast. Comparing to conventional EJB based architecture, spring is much easier to use and deploy, and it has the trend to take the position of EJB in J2EE. Not long ago, Adobe published Flex 2.0, of which the most important part is Flex Data Service. Flex Data Service is similar to Web component or EJB component in J2EE, can be deployed in J2EE application server, supplying RPC service, message service and data management service, very powerful. Bond Trading System is constructed basing on Flex Data Service.This paper take Bond Trading System as an example, discusses about the issues may face during the development of Rich-Client financial trading system, including technique selection, project design, system performance, and potential problems. It supplies some advice to the construction of Rich-Client application. Firstly it describes the research background, including rich client technology and Spring light weight J2EE develop framework. Then emphasize to introduce the Flex rich client technology, including three parts content of Flex Data Service: RPC service, message service and data management service, and client MVC framework —Cairngorm. About Bond Trading System, is introduced by four parts: business requirement, overall architecture, server side implementation, client side implementation, and list out some functional module's implementation and relative technologies. System was tested by Flex stress test framework, list and analysis some of the test result, and find the bottleneck of system performance. At last, make some analyze on system's real time ability and reliability, find some potential problems, then give out a way to resolve the problems—integrating with mature JMS product and making distribute transaction management, and gives out the implementation.