MRI-R2 Consortium: Development of Dynamic Network System (DYNES)

点击翻译按钮获取中文摘要

This award is funded under the American Recovery and Reinvestment Act of 2009 (Public Law 111-5). This project will develop and deploy the Dynamic Network System (DYNES), a nationwide cyber-instrument spanning 39 US universities and 16 regional networks. DYNES will support large, long-distance scientific data flows in the LHC, other leading programs in data intensive science (such as LIGO, Virtual Observatory, and other large scale sky surveys), and the broader scientific community. By integrating existing and emerging protocols and software for dynamic circuit provisioning and scheduling, in-depth end-to-end network path and end-system monitoring, and higher level services for management on a national scale, DYNES will allocate and schedule channels with bandwidth guarantees to several classes of prioritized data flows with known bandwidth requirements, and to the largest high priority data flows, enabling scientists to utilize and share network resources effectively. DYNES is dimensioned to support many data transfers which require aggregate network throughputs between sites of 1-20 Gbps, rising to the 40-100 Gbps range. This capacity will enhance researchers? ability to distribute, process, access, and collaboratively analyze 1 to 100 TB datasets at university-based Tier2 and Tier3 centers now, and PB-scale datasets once the LHC begins operation.DYNES is based on a ?hybrid? packet and circuit architecture composed of Internet2's ION service and extensions over regional and state networks to US campuses. It will connect with transoceanic (IRNC, USLHCNet), European (GÉANT), Asian (SINET3) and Latin American (RNP and ANSP) Research and Education networks. It will build on existing key open source software components that have already been individually field-tested and hardened: DCN Software Suite (OSCARS / DRAGON), perfSONAR, UltraLight Linux kernel, FDT, FDT/dCache, FDT/Hadoop, and PLaNeTs. The DYNES team will partner with the LHC and astrophysics communities, OSG, and Worldwide LHC Computing Grid (WLCG) to deliver these capabilities to the LHC experiment as well as others such as LIGO, VO and eVLBI programs, broadening existing Grid computing systems by promoting the network to a reliable, high performance, actively managed component. Future science programs in HEP, astrophysics and gravity wave physics, and other data intensive disciplines, will be facilitated by DYNES? technologies and worldwide network partnerships. Working with CHEPREO and similar education and outreach efforts targeting under-served communities both in the US and overseas, DYNES will reach a wide variety of students at collaborating institutes including underrepresented groups and minorities. This will lower the barriers, and enable individual graduate students, undergrads, postdocs and faculty to use DYNES to achieve high throughput in support of their research in many data intensive fields.