US-Federal Republic of Germany Cooperative Research: (Computer Science) Rapid prototyping for Distributed Applications

该奖项支持海军研究生院的陆奇教授和Berzins教授以及一名博士后前往德国国家计算机科学研究中心（GMD）系统工程研究所进行研究访问，他们将主要与Bernd J. Kraemer博士合作。他们提出的研究合作的目标是扩展适用于硬实时系统的快速原型语言PSDL，以包括适合分布式应用程序的规范概念和实现技术。该计划的几个主要组成部分将是开发一个语义模型，支持分布式程序的原型和可用的验证技术，在资源争用的情况下，与多处理器上的原型执行相关的调度和资源分配问题。还有一种两级的原型执行方法，它使用可执行的规范和可重用的软件组件。该项目计划依赖于合作各方在语言和工具设计、实验和理论分析方面的部分互补技能，以减少在解决共同关心的问题时使用的资源。世界上这一领域的一些最好的工作是在德国完成的，美国的高级调查人员受到高度重视。寻找有效和高效的方法来确定和验证软件系统的需求是软件工程中未解决的重要问题。原型是一个迭代的开发和演变过程，它承诺产生可靠的和用户接受的软件。多处理器系统的出现对软件规格、设计和验证提出了新的挑战。正确性检查和死锁检测变得至关重要，因为分布式程序中常见的不确定性和无限行为使得无法对它们进行充分的测试。这里提出的研究将对这些问题作出切实的贡献。

This award supports Professors Qi Lu and Berzins of the Naval Postgraduate School, plus a postdoctoral associate, for research visits to the Institute for Systems Engineering of the German National Research Center for Computer Science (GMD), where they will collaborate primarily with Dr. Bernd J. Kraemer. The objective of their proposed research collaboration is to extend the rapid prototyping language PSDL, which applies to hard, real-time systems, to include specification concepts and implementation techniques suitable for distributed applications. Several major components of the plan will be development of a semantic model supporting both prototyping and available verification techniques for distributed programs, scheduling and resource allocation issues related to prototype execution on multiple processors in the presence of resource contention, and a two level approach to prototype execution which uses both executable specifications and reusable software components. The project plan relies on partly complementary skills in language and tool design, experimentation, and theoretical analysis of the cooperating parties to reduce the use of resources in solving problems of common concern. Some of the best work in this area in the world is being done in Germany, and the U.S. senior investigators are highly regarded. Finding effective and efficient methods for determining and validating the requirements for a software system is an important unsolved problem in software engineering. Prototyping is an iterative development and evolution process that promises to produce reliable and user-accepted software. The emergence of multiprocessor systems poses new challenges to software specifica- tion, design, and validation. Correctness checking and deadlock detection become crucial because the nondeterminism and infinite behaviors commonly found in distributed programs make it impossible to test them sufficiently. The research proposed here will make a solid contribution to these problems.