CC-NIE Networking Infrastructure: High Speed Science DMZ @ Johns Hopkins for Data Intensive Computing

该网络基础设施项目的重点是创建一个先进的专用研究网络，以促进大型科学数据集的传输，同时消除传统的基于数据包的检测技术的障碍。该项目的目标有三个：1)建立一个强大的以研究为导向的科学DMZ，简化网络边界；2)扩展到整个机构内部研究小组的高速连接；3)连接到2014年开放的新建高性能研究计算设施（HPRCF）。快速将大型科学数据集从源头转移到目的地的能力至关重要。该项目的主要目标是创建一个新的网络边界，以显着提高大规模数据集的传输速率。在网络吞吐量达到100G的情况下，传统的基于包检测的防火墙技术延时过大，需要消除。Science DMZ将采用较新的网络协议，例如虚拟设备上下文和多包层交换，以实现更快的融合。JHU是几个数据密集型活动的中心，这些活动正在改变全国各地的整个社区。世界上最常用的天文设施——斯隆数字巡天的档案，托管在JHU，我们的100TB的公共湍流数据库在很大程度上负责数值模拟的新访问范式。新的HPRCF将在JHU和马里兰大学的科学家之间建立新的合作关系。该项目将改造JHU骨干网，为更广泛的社区带来高速连接。

This cyberinfrastucture project is focused on creating an advanced dedicated research network that will facilitate the transfer of large scientific datasets while eliminating the barriers of traditional packet based inspection technologies. The goal of this project is threefold: 1) to establish a robust research oriented Science DMZ that simplifies the network perimeter, 2) to expand high-speed connectivity to internal research groups throughout the institution, and 3) to connect to a newly constructed High Performance Research Computing Facility (HPRCF) opening in 2014.The ability to quickly transfer large scientific data sets from origination to destination is of paramount importance. The principal goal of this project will be the creation of a new network perimeter to significantly increase the transfer rates of large scale datasets. With network throughput speeds reaching 100G, traditional firewall technologies relying on packet based inspection introduce too much latency and need to be eliminated. The Science DMZ will embrace newer network protocols, such as Virtual Device Contexts and multi-packet layer switching to enable faster convergence.JHU is at the center of several data-intensive activities that are transforming whole communities across the country. The archive of the Sloan Digital Sky Survey, the world's most used astronomy facility, is hosted at JHU, and our 100TB public turbulence database is largely responsible for a new access paradigm to numerical simulations. The new HPRCF will create new collaborations between JHU and University of Maryland scientists. This project will transform the JHU backbone and bring high-speed connectivity to a much broader community.