U.S.-France Cooperative Research (INRIA): Theory and Implementations of Strongly Consistent Distributed Shared Memory Systems.

这个为期三年的奖项支持美国和法国在计算机操作系统和系统软件方面的合作研究。研究人员是Masaaki Mizuno，堪萨斯州立大学的Gurdip Singh，俄克拉荷马州立大学的Mitchell Neilsen以及雷恩大学和INRIA（法国国家计算机科学与应用数学研究所）的Michel Raynal。他们的研究目标是研究分布式共享内存系统的定义和实现，通过开发一个通用的框架来分类协议和评估有效的实现。他们将开发适用于不同计算机体系结构的协议。美国调查人员将数学分析、计算机体系结构和软件实现方面的专业知识带入此次合作。这是由法国在分布式计算的广泛领域的专业知识补充的。该项目还利用了INRIA的大量设备，包括将用于实施研究的分布式并行机（具有32个处理器的Hypercube）。分布式共享内存系统很重要，因为它们简化了编程。该项目将进一步了解这些系统，其结果可应用于高性能计算环境中的数据库和分布式文件系统的复制。* * *

9406785 Mizuno This three-year award supports U.S.-France cooperative research in computer operating systems and system software. The investigators are Masaaki Mizuno, Gurdip Singh of Kansas State University, Mitchell Neilsen, Oklahoma State University and Michel Raynal of the University of Rennes and INRIA (French National Institute for Research in Computer Science and Applied Mathematics). The objective of their research is to study the definitions and implementations of distributed shared memory systems by developing a general framework for classifying protocols and for evaluating efficient implementations. They will develop protocols suited to different computer architectures. The U.S. investigators bring to this collaboration expertise in mathematical analysis, computer architectures, and software implementation. This is complemented by French expertise in broad areas of distributed computing. The project also takes advantage of a large array of equipment at INRIA, including a distributed parallel machine (a Hypercube with 32 processors), which will be used in the implementation studies. Distributed shared memory systems are important because they simplify programming. This project will further understanding of these systems and its results could be applied to replication of databases and distributed file systems in a high performance computing environment. ***