一种混合模式两级抽象服务粒度划分法

该文首先介绍了当前服务粒度划分研究现状，并对几种主流服务粒度划分方法进行分析，指出了它们的不足之处，进而提出了一种基于业务、组件、类的混合模式两级抽象粒度划分方法——TAMM，该方法利用混合模式并结合java设计模式思想对服务粒度划分过程进行两级优化组合抽象，以适中的粗细粒度优化组合抽象服务，从而提高服务的稳定性、灵活性及性能。最后，结合实际案例对该划分方法的特性进行了分析说明。

A Mingle Pattern to Divide the Service Granularity in Twice Abstract

The article firstly reviewed the previous research of service granularity, then analyzed the prevailing methods of its division and pointed out the drawbacks. What＇s more, the author further proposed the mingle pattern that is a division method of twice abstract granularity based on the business, component and class. This method made use of the mingle patterns combined with the conception of JAVA Design Patterns to optimize and abstract of the procedure of dividing the service granularity. With proper size of granularity to optimize and abstract of the service, this pattern may improve the stability, flexibility and function of the service. Finally, the author also illustrated the function of this pattern by employing some practical cases.