CC*DNI Networking Infrastructure: Science DMZ and SDN

CC*DNI项目通过建立两个连接到爱达荷州区域光网络(FIRE)的科学DMZ来扩展博伊西州立大学的研究网络。科学DMZ位于大学网络的边缘，位于爱达荷州计算联盟的网站上，托管在爱达荷州国家实验室超级计算中心。博伊西州立大学利用软件定义网络(SDN)、OpenFlow交换机和OpenFlow控制器，使Iron能够在大学研究人员和国家实验室的高性能计算(HPC)设备之间以高吞吐速率交换数据流量。这一增强的区域研究网络为我们的研究人员正在使用大学、国家科学基金会和能源部资产进行并行计算分析的大型数据集提供了快速、极其安全和动态的网络环境。在整个区域光网络中使用监控软件PerfSONAR框架允许持续监控吞吐量、性能和延迟问题。随着数据的范围和数量的增加，以及博伊西州立大学对分布式和国家网络基础设施资产的日益依赖，提高传输大型数据集的控制、速度和便利性对于推进博伊西州立大学的研究至关重要。SDN交换机和控制器允许跨区域光网络的高传输速度，但由于路由流量的延迟，传输速度通常较慢。可靠的高速数据流量是研究人员在扩大大规模数据密集型科学研究时所期望的基础设施。该项目增强了整个爱达荷州研究社区、落基山高级计算机联盟和XSEDE的高性能计算。

The CC*DNI project expands Boise State University's research network by establishing two Science DMZs connected to the Idaho Regional Optical Network (IRON). The Science DMZs are located on the edge of the University network and at the Idaho Computing Consortium site, hosted at the Idaho National Laboratory Supercomputing Center. Utilizing Software Defined Networks (SDN), OpenFlow Switches and an OpenFlow Controller, Boise State enables IRON to switch data traffic at a high throughput rate between researchers at the University and the High Performance Computing (HPC) equipment at the National Laboratory. This enhanced regional research network provides a fast, extremely secure and dynamic network environment for the larger data sets our researchers are analyzing with parallel computing using University, National Science Foundation and Department of Energy assets.Utilizing the monitoring software perfSONAR framework throughout the regional optical network allows constant monitoring of throughput, performance and latency issues. Improving the control, speed and ease of transferring large data sets is imperative to advancing Boise State research as the scope and volume of data increases and as the University increasingly relies on distributed and national Cyberinfrastructure assets. The SDN switches and controller allow high transport speed across the regional optical network that is normally slower due to the latency of routed traffic. Reliable high speed data traffic is the infrastructure that researchers expect as they expand large scale data intensive scientific research. This project enhances the high performance computing throughout the Idaho research community, the Rocky Mountain Advanced Computer Consortium and XSEDE.