MRI: Acquisition of Infrastructure for Research in Grid Computing and Multiscale Systems Computation

该项目将采购设备，形成伦斯勒网格的核心，这是一个与未来国家网络基础设施能力相一致的研究试验台，支持信息技术、生物技术、纳米技术和国土安全方面的研究项目。项目重点是网格中间件和多尺度计算技术的发展，以及这些技术在科学和工程中的应用，推进和使用网格。RPI核心网格由3个集群组成，共有374个处理器，第4个集群有162个处理器，通过专用的10Gbit链路连接在一起。由此产生的网格为具有异构连接的异构集群提供了一个研究测试平台，可以为应用程序提供大量的处理能力。计划中的网络基础设施（由分布式异构计算集群组成）将重点放在有效使用所需的中间件和仿真技术的开发上，为工业、医药和商业提供计算能力。网格计算技术研究包括分布式编程技术开发、网格网络仿真与安全、数据流、提取与控制等项目；动态负载平衡，支持实时应用程序和设备。多尺度计算应用包括自适应多尺度计算工具（支持自适应动态仿真模型）；生物分子结构的多尺度方法；基于仿真的心血管疾病医疗规划多功能高分子纳米复合材料；动态联系任务规划及其在灾难恢复中的应用以及通信网络中社会群体的进化。鼓励与现有网络基础设施投资的合作努力，包括NSF PACI中心和ETF，这一校园级的努力涉及来自11个科学和工程院系的38名PI教师，涉及99名学生。这项基础研究项目得到了7个联邦机构、纽约州和7家公司/基金会的资助，包括13所合作大学、6个国家实验室和一个州机构。更广泛的影响包括：通过开放标准和开源互操作工具使进步广泛应用；培训和参与130多名校园研究人员，以及其他大学、实验室和行业的合作同事；通过完善的技术转移计划将技术直接转移到行业，包括初创企业；扩大教育计划，包括与网络基础设施和多尺度系统技术相关的新课程；将相关项目纳入现有课程，并向不同背景的大学预科学生演示基于网格的模拟。

This project, acquiring equipment to form the core of the Rensselaer Grid, a research testbed consistent with the capabilities of the future national cyberinfrastructure, supports research programs in Information Technology, Biotechnology, nanotechnology, and Homeland Security. Projects focused on the development of grid middleware and multiscale computing technologies, and the application of these technologies to science and engineering, advance and use the grid. The RPI core grid consists of 3 clusters having a total of 374 processors, and a 4th cluster with 162 processors connected via dedicated 10Gbit links. The resulting grid provides a research testbed of heterogeneous clusters with heterogeneous connections that can provide substantial processing power to applications. Placing emphasis on the development of middleware and simulation technologies needed to make effective use, the planned cyberinfrastructure, consisting of distributed heterogeneous computing clusters, provides computing power to industry, medicine, and business. The grid computing technologies research consists of a set of projects including the development of distributed programming technologies, grid networking simulation and security, data flow, extraction and control; dynamic load balancing, and support of real time applications and devices. The multiscale computation applications include adaptive multiscale computation tools (supporting adaptive dynamic simulation models); multiscale methods for biomolecular structure; simulation based medical planning for cardiovascular disease; multifunction polymeric nanocomposites; dynamic contact tasks planning with application to disaster recovery; and the evolution of social groups in communication networks. Encouraging collaborative efforts with existing cyberinfrastructure investments, including NSF PACI centers and the ETF, this campus level effort engages 38 PI faculty from 11 academic departments in science and engineering and involves 99 students. The underlying research project, supported by scores of grants from 7 federal agencies, NY state, and 7 companies/foundations, include 13 collaborative universities, 6 national labs, and a state agency. Broader Impacts involve: Making the advances widely available through open standards and open source interoperable tools, Training and involvement of over 130 on-campus researchers, along with collaborating colleagues in other universities, labs, and industry, Transferring direct technology to industry, including start-ups, through well established technology transfer programs, Expanding the educational programs, including new courses related to cyberinfrastructure and multiscale systems techniques, and inclusion of relevant projects in existing courses, and Demonstrating grid-based simulations to precollege students of diverse background.