CC*IIE Networking Infrastructure: FLowell: Accelerating Data-Driven Scientific Research at the University of Massachusetts Lowell

Flowell项目通过软件定义的10 Gbps科学网络增强校园网络基础设施（CI），在马萨诸塞州绿色高性能计算中心（MGHPCC）部署先进的全网络（SDN）集成计算/存储机架，并以10 Gbps的速率将研究人员实验室连接到MGHPCC以及外部合作者，以实现高效的数据共享和传输。UML校园已经看到了许多学科的数据密集型研究的巨大增长，如大气科学，网络科学，生物医学科学和数据管理。为了满足这些项目的数据传输需求，需要解决网络访问速度慢、网络管理策略受限以及与MGHPCC和其他外部计算/存储资源连接效率低下的问题。UML校园网采用软件定义网络技术和专用的科学隔离区（DMZ）概念，消除了传输大型数据集的瓶颈。第2层软件定义的控制、基于测量的perfSONAR（性能面向服务的网络监控架构）和数据传输节点被用来实现最先进的网络控制和操作。Flowell是一个优秀的网络环境，PI评估和利用新的网络技术，并欢迎来自其他研究人员更广泛的参与。基于PI和校园IT之间的持续合作以及OpenFlow交换机的试点部署，科学团队，IT部门，供应商和MGHPCC人员形成了强大的合作伙伴关系，以建立先进的基础设施，并在全球范围内加速以数据为中心的知识发现。

The FLowell project accelerates a set of data-driven science projects at the University of Massachusetts Lowell (UML) by enhancing campus cyberinfrastructure (CI) with a software defined 10Gbps science network, deploying an advanced and fully network (SDN) integrated computing/storage rack at the Massachusetts Green High Performance Computing Center (MGHPCC), and connecting researchers labs at 10Gbps rate to the MGHPCC as well as external collaborators for highly efficient data sharing and transfer. The UML campus has seen a tremendous growth of data intensive research in many disciplines such as atmospheric science, network science, biomedical science and data management. To fulfill the data transfer needs from these projects, there is a strong call to address the slow network access speed, constraints on network management policies, and inefficient connectivity to MGHPCC and other external computing/storage resources. The UML campus network embraces software defined network technologies and the dedicated science Demilitarized Zone (DMZ) concept, which remove the bottlenecks for transferring large datasets. Layer 2 software defined control, perfSONAR (PERFormance Service Oriented Network monitoring ARchitecture) based measurement, and data transfer nodes are leveraged to enable state-of-the-art network control and operations. FLowell is an excellent network environment for PIs to evaluate and take advantage of novel networking technologies and welcomes a broader participation from other researchers. Based on the ongoing collaboration between the PI and campus IT and the pilot deployment of OpenFlow switches, the science teams, IT department, vendors, and MGHPCC personnel form a strong partnership to establish the advanced infrastructure and accelerate data centric knowledge discovery in a global context.