论文部分内容阅读
透明计算系统是一种支持程序按需加载的新型分布式系统。缓存机制是影响透明计算系统性能重要因素之一。针对现有的缓存仿真算法不支持透明计算缓存性能分析的问题,该文提出了一种支持多用户共享服务器存储空间的缓存仿真算法(SSCS)。该方法扩展了传统的栈距离模型,采用首存储块号、拥有者用户ID二元组标识缓存块,根据不同用户对缓存块的访问情况分别管理、调度和测量。实验表明,基于SSCS的仿真结果可用于指导透明计算系统的缓存设计。根据仿真结果进行改进,可以有效提升透明计算系统的性能和可扩展性。
Transparent computing system is a new type of distributed system that supports the loading of programs on demand. Caching mechanism is one of the important factors affecting the performance of transparent computing system. Aiming at the problem that the existing caching simulation algorithm does not support the performance analysis of transparent computational cache, this paper proposes a caching simulation algorithm (SSCS) that supports the storage space of multi-user shared server. This method expands the traditional stack distance model, identifies the cache block by using the first storage block number and the owner user ID two-tuple, and separately manages, schedules and measures the access of the cache block according to different users. Experiments show that SSCS-based simulation results can be used to guide the transparent computing system cache design. According to the simulation results to be improved, can effectively improve the performance and scalability of transparent computing system.