CC*IIE Networking Infrastructure: University of California Riverside's Science DMZ

在加州大学河滨分校(UCR)，科学家们正在使用一种被称为“科学非军事区”或DMZ的高速网络，以保持他们的海量科学数据在校园内外的合作者之间流动。UCR-DMZ由美国国家科学基金会资助，是一个私人的？高带宽网络，支持(I)新的协作方式；(Ii)以更高的分辨率收集数据；(Iii)全新的数据密集型项目；(Iv)更多的大数据项目，包括学生研究。UCR-DMZ使用空气吹制光纤(ABF)光纤--字面意思是通过管道吹制的高带宽光纤--允许以低成本增加未来的容量。该网络为加州大学S分校的网络基础设施做好了准备，预计未来五年教职员工人数将增长40%，同时每位教职员工的数据也将增长。UCR-DMZ支持数据密集型研究，包括实时监测土壤生物活动的地下摄像机、基于翼拍数据的昆虫识别、宇宙引力模型的传播、视频生物信息学、下一代DNA测序、互联网性能预测和可持续能源项目。研究数据正在由加州大学S图书馆存档。对于许多UCR实验室来说，TB级的数据集越来越常见。UCR-DMZ提供1)致力于研究的“无摩擦”网络路径，不与日常商业和企业计算竞争；2)高性能数据传输节点(DTN)，具有端到端优化的数据传输仪器、硬件和操作系统；以及3)在面向性能服务的网络监控体系结构(PerfSONAR)中使用测量/测试节点进行性能监控。

At the University of California, Riverside (UCR), scientists are using a high-speed network known as a "science demilitarized zone" or DMZ to keep their massive science data flowing between collaborators on and off campus. Funded by the NSF, the UCR-DMZ is a ?private? high-bandwidth network that enables (i) new ways of collaborating; (ii) collection of data at much higher resolution; (iii) entirely new data-intensive projects; (iv) a much larger number of Big Data projects, including student research. The UCR-DMZ uses Air Blown Fiber (ABF) optical fiber -- high-bandwidth strands literally blown through conduits -- that allow low-cost addition of future capacity. The network prepares UCR?s Cyberinfrastructure for an anticipated 40% growth in faculty in the next five years, along with growth in data-per-faculty-member. The UCR-DMZ supports data-intensive research that includes underground cameras to monitor soil biological activity in real time, insect identification based on wing-beat data, dissemination of gravitational models of the Universe, video bioinformatics, next generation DNA sequencing, Internet performance prediction, and sustainable energy projects. Research data is being archived by UCR?s libraries. Terabyte datasets are increasingly routine for many UCR labs. UCR-DMZ offers a 1) a "friction free" network path devoted to research, not competing with day-to-day business and enterprise computing; 2) high-performance Data Transfer Nodes (DTNs) with instrumentation, hardware and operating system optimized end-to-end for data transfer; and 3) performance monitoring using measurement/test nodes in a PERFormance Service Oriented Network monitoring Architecture (perfSONAR).