GLORIAD: Global Ring Network for Advanced Applications Development
The Global Ring Network for Advanced Applications Development (GLORIAD) aims to better integrate, with an advanced network infrastructure, the science and education (S&E) communities of the US, Russia, China, Korea, Canada and Netherlands - as well as broader Europe as well as Eastern and Central Asia. It includes: a jointly managed, hybrid circuit-switched and routed network designed in a ring topology encircling the northern hemisphere; and services to support increased S&E cooperation for both general and highly advanced (in terms of network needs) user communities. In addition to supporting active scientific exchange with network services, the project will provide a production environment for advancing the state-of-the-art in collaborative and network technologies, including Grid-based applications development, end-to-end circuit provisioning, IPv4/IPv6 an d unicast/multicast services, advanced network monitoring, performance measurement and network security research. Supporting the increased growth in global S&E cooperation, GLORIAD is actively cooperating with the world's most advanced S&E networks and programs to create a rational, coherent strategy and architecture for the future of S&E networking.
GLORIAD concluded its NSF-sponsored award under the IRNC Program (2005-2012) establishing with its international partners multiple network services (many at 10 Gbps) ringing the northern hemisphere and dedicated to advancing international research and education (R&E) collaboration. Full report and all associated appendices detailing usage at: www.gloriad.org/gloriaddrupal/nsf_report. Significant outcomes: GLORIAD was the first network to connect US and China R&E communities; 10 Gbps hybrid (circuit- and packet-switched) network (including production Layer-3 routed IP service); Substantial growth: total traffic volume of 7.4 petabytes; 5.6 petabytes during 2005-2012; and 1.3 petabytes additional through April 2013. Unique user IP addresses grew from <1 million to 12+ million. Almost 90 countries served as source and nearly 100 as destination traffic. 8 new partner countries; Upgraded network services with production quality service for specialized networks, and dedicated capacity for large applications; Working prototype of dvNOC (distributed, virtual network operations center), integrating utilization, performance, operations and security data into a web-based environment for global operations teams to manage network services; System constantly monitors traffic for "bad actor" IP addresses, heavy increase in packets and asnums; Identified and solved performance issues with ESNet-CSTNet; all L3 traffic monitored with passive and active monitoring systems to proactively address end-to-end performance problems; Databases: performance/utilization and usage case studies, with GLORIADEarth displaying current top users and updated every 10 seconds; GLORIAD Classroom: community-building clearinghouse on cyberinfrastructure, users, applications, techniques. Security: By 2009, we installed an array of security monitoring equipment and methodologies to improve network security. Network GLORIAD rings 29,000 miles around the northern hemisphere and was the first to connect US-China R&E networks. GLORIAD crosses North America from StarLight/Chicago to NYC via CANARIE. New partners: NLR brought 10G capacity, and a 1 GbE circuit was engineered for GLORIAD between SouthernLightRail (SLR) PoP in Atlanta and GLORIAD PoP in Chicago. GLORIAD crosses the Atlantic via Tata, SURFnet and IRNC TransLight, connecting at Netherlight in Amsterdam to MoscowLight via Russia-provided 10G circuits and NORDUnet; across European Russia to Novosibirsk and Khabarovsk exchange points, and to Hong Kong Open Exchange Point (HKLight). GLORIAD takes redundant paths via Tata to Seattle. One, now 10 Gbps, directly from Hong Kong to Seattle, supports US-Chinese Academy of Sciences. The other (now) 25 Gbps hybrid service connects between Daejeon, South Korea and Seattle. Partnerships GLORIAD added 8 countries to original three partners (US/Russia/China): Canada, Korea, Netherlands and 5 Nordic partners and, subsequently, India, Egypt and Singapore. GLORIAD supports "science diplomacy" to foster mutual understanding. dvNOC In 2008, US, China and Korea partners launched the dvNOC for integrating utilization, performance, operations and security data into a shared web-based environment for serving broad access and interoperability of R&E networks. Network Usage GLORIAD’s automated queries report "top users," applications, traffic volume, packet loss. Top users include HEP, astronomy, atmospheric sciences, health sciences, biological sciences, nuclear materials protection, with FermiLab as #1 user. 750 US colleges/universities, 40 federal agencies, 10 national laboratories, connecting to 10,000+ science institutions globally use GLORIAD. Our data confirms substantial growth of network size and utilization. Total historical traffic volume is 7.4 petabytes; 5.6 petabytes during 2005-2012; and 1.3 petabytes more through April 2013. Between 2005-2010, the number of unique user IP addresses jumped from <1 million to 12+ million. Almost 90 countries served as the source of traffic routed across GLORIAD and nearly 100 received data. GLORIAD’s broad impact is demonstrated by the growth and high volume of utilization. By 2010, GLORIAD Earth’s monitoring system – a dvNOC application – was enhanced to provide new metrics, showing all shared IP service flows transiting Chicago (and now Seattle), and displaying top 25 users, every 10 seconds (www.gloriad.org/dvnoc, "GLORIAD Earth" tab). A new (2013) detailed report can be found at this link: www.gloriad.org/gloriaddrupal/nsf_report Databases Two databases were developed to better understand user activity – enabling better service to new/existing users, special reports, and information to broader public: Usage/Performance, and Case Studies for GLORIAD Classroom, a community clearinghouse about R&E infrastructure, users, applications, best practices. Security To address cybersecurity, we implemented capabilities to detect and alert on network use and anomalies and promote wide-area security as an open and globally federated effort — facilitating information exchange both vertically (down to the institution level) and horizontally (across countries and regions). By 2009, we installed a wide array of security monitoring equipment and methodologies: • All systems "kerberized" for safe, secure authentication and access. • RT system maintains tickets for managing anomalous network use (high retransmits, outages, capacity outside of "normal" levels, new institutions and AS numbers, network use by institutions within "export controlled" countries, "bad actors"). • perfSONAR installed in Chicago, Seattle. • Nagios alerts to problems on all GLORIAD equipment. • Cacti-generated graphics of overall circuit and VLAN traffic levels.
