Distributed Cache Pattern Research And Apllication In Financial System

Distributed cache is a key technology for the modern enterprise system to gain the high performance and high scalability. Distributed cache technology is popular, but most of the researches are focus on the field of cache technology for web caching. While this paper is mainly about the characteristics and applications of distributed cache technology in enterprise systems. Different systems have different designs of distributed cache framework. There is no general distributed cache framework. This paper summarizes the distributed cache patterns based on the analysis of the current cache technology. The purpose is let these patterns easy the future implementation of distributed cache technology. These patterns include data distribution pattern, data synchronization pattern and cache replacement algorithm, etc.Since financial service is replete with "delightfully parallel" apps and large volume of data, distributed cache is especially popular and important in financial application. Take a financial system - CX for example, this paper describe how to apply the distributed cache patterns to CX, including the detail design and implementation. CX is a peer-to-peer distributed system; it provides a foreign exchange trading platform. Users want to trade at the latest rate, so it's undesirable to trade at the expired rate due to the slow response of CX system. Thus it's important to use distributed cache technology to improve the performance of CX. Appropriate distributed cache patterns should be chosen according to the architecture of CX, e.g. Client invalidation pattern is one of the right choices. The distributed cache framework of CX mainly consists of four parts: Cache API, Cache Map, Cache Synchronization and Dependency Check. The performance test of CX system can prove that distributed cache technology does enable CX system gain better performance.