论文部分内容阅读
为了满足面向服务体系架构中对大规模访问请求高效评估的需求,提出了一种分布式策略评估引擎XDEngine.该引擎的核心为策略分解模块和请求分配模块,具有分解策略和分发请求的功能.提出了按照主体属性进行策略分解的算法,该算法将策略分解成多个子策略,并使得部署到每个策略决策点的子策略的开销相等或近似相等.通过实验对比了XDEngine和Sun PDP中策略决策点的评估性能,测量了含有不同数量策略决策点的XDEngine的评估时间.实验结果表明:策略分解算法大幅度提高了XDEngine的评估性能,XDEngine的评估时间随着策略决策点数目的增多而降低;XDEngine能够满足面向服务体系架构中高效评估高并发量访问请求的需求.
In order to satisfy the demand for efficient evaluation of large-scale access requests in service-oriented architecture, a distributed policy evaluation engine XDEngine is proposed.The core of this engine is the strategy decomposition module and request distribution module, and has the function of decomposing policies and distributing requests. This paper proposes an algorithm that decomposes policies into multiple sub-policies according to the principal attributes and equal or approximately equal the costs of the sub-policies deployed to each policy-decision point.Through experiments comparing the strategies in XDEngine and Sun PDP The evaluation performance of XDEngine is evaluated, and the evaluation time of XDEngine with different number of policy decision points is measured.The experimental results show that the strategy decomposition algorithm greatly improves the evaluation performance of XDEngine, and the evaluation time of XDEngine decreases as the number of policy decision points increases; XDEngine addresses the need to efficiently evaluate high-concurrent requests for access in service-oriented architectures.