CISE Research Resources: Infrastructure for Research in Parallel and Distributed Computing

EIA 0224369 Pandey，RajuAksoy，Demet；Devanbu，Premkimar T.；Olsson，Ronald A.加州大学戴维斯分校：研究并行和分布式计算的基础设施这项提案，支持并行计算、分布式系统、无处不在(自主)系统和卫星通信，旨在在软件系统中心内建立一个并行和分布式计算实验室。该实验室将包括四个集群，每个集群通过高速网络连接。并行计算集群将托管多个多处理器系统；分布式系统集群将包括工作站和个人计算机，泛在系统集群将连接嵌入式设备(包括手持设备传感器)；卫星集群将托管发起下行通信以传播数据的设备。传感器网络将主要通过专用无线网络连接，该网络将被路由到主要的高速主干。该基础设施将满足1中四个研究项目的需要。分布式系统，2.并行计算，3.分布式系统软件工程，4.广域数据传播。分布式系统集群将使分布式系统组能够开发高效、可扩展、健壮和安全的分布式系统，以适应底层系统条件的变化。此外，无处不在的系统集群将允许该集团开发新的特定于应用的路由算法、新的编程模型可定制的中间件无缝集成基于传感器的系统和可扩展的资源分配，以及用于大型基于传感器的系统的保护算法。卫星通信集群和分布式系统集群将使分布式数据传播小组能够开发高度可伸缩、稳健和安全的数据传播技术，并提供可调整的服务质量。并行编程小组将使用并行计算集群来评估他们在并行编程语言和并行系统的高效运行时系统方面的工作。在教育方面，基础设施将在几门课程中用作实验试验台，并将形成基础，为研究型学生提供并行和分布式系统的开发和分析技能。

EIA 0224369Pandey, RajuAksoy, Demet; Devanbu, Premkimar T.; Olsson, Ronald A.University of California - Davis CISE RR: Infrastructure for Research in Parallel and Distributed Computing This proposal, supporting parallel computing, distributed systems, ubiquitous (autonomous) systems, and satellite communication, aims at establishing a parallel and distributed computing laboratory within the center for software systems. The lab will include four clusters, each connected through high speed network. The parallel computing cluster will host several multiprocessor systems; the distributed system cluster will include workstations and PCs, the ubiquitous systems cluster will connect embedded devices (including handheld device sensors); and the satellite cluster will host equipment for initiating the downlink communication for disseminating the data. The sensor network will primarily be connected through an ad hoc wireless network that will be routed to the main high-speed backbone. The infrastructure will meet the needs of four research projects in1. Distributed Systems,2. Parallel Computing,3. Distributed Systems Software Engineering, and4. Wide Area Data Dissemination.The distributed systems cluster will enable the distributed systems group to develop highly efficient, scalable, robust, and secure distributed systems that adapt to changes underlying systems conditions. Further, the ubiquitous systems cluster will permit the group to develop new application-specific routing algorithms, novel programming models customizable middleware that seamlessly integrate sensor-based systems and scalable resource allocation, and conservation algorithms for large sensor-based systems. The satellite communication cluster, along with the distributed systems cluster, will enable the distributed data dissemination group to develop highly scalable, robust, and secure data-dissemination techniques with adaptable quality of service. The concurrent programming group will use the parallel computing cluster to evaluate their work on both concurrent programming languages and efficient runtime systems for parallel systems. On the educational side, the infrastructure will be used as experimental test-beds in several courses and will form the basis for providing research students with development and analysis skills in parallel and distributed systems.