- Home
- Advanced Services and Applications
- Large Scale eScience Networking
Large Scale eScience Networking
One key focus for iCAIR and its research partners is developing communication services and networks for large-scale, high-performance e-Science applications, which have always been key drivers of next-generation technologies. These highly distributed, resource-intensive initiatives often encounter technology barriers years before other communities recognize them.
Many eScience applications are data, bandwidth, compute, and storage intensive. Research science communities employ large-scale resources distributed worldwide, including instrumentation (e.g., synchrotrons, radio telescopes, sensors observatories), data repositories, HPC (supercomputing for computational science), and analytic sites (e.g., AI/ML/DP) sites, that have to be interconnected into common ecosystems that can support high capacity, high-performance workflows.
Such research communities include those investigating topics in high-energy physics, astrophysics, bioinformatics, computational biology, genomics, fusion energy, computational chemistry, nanotechnology, data mining, digital engineering, geosciences, oceanographic and atmospheric studies, space exploration, medical imaging, materials science, and advanced digital media, including ultra-high-resolution visualization, specialized virtual-reality, mixed reality, tele-immersion, 3D, and 8k media.
High Energy Physics (HEP)
For many years, iCAIR has been engaged in multiple cooperative initiatives with the worldwide high-energy physics research (HEP) community. This project focuses on problems related to managing, transporting, and storing extremely large amounts of HEP data. HEP research investigates complex topics related to the fundamental nature of matter, especially the attributes and behavior of the smallest elemental particles.
These scientific investigations are undertaken by collaborative research communities worldwide that use highly sophisticated instrumentation to gather extremely large amounts of data, which then is distributed for analysis worldwide. Primary HEP reference projects include the Large Hadron Collider at CERN in Geneva, Switzerland, which generates more data than any other science project worldwide.
This data is transported to compute sites worldwide, including primary sites (Tier 1) and secondary (Tier 2 and Tier 3), on two major international networks, the Large Hadron Collider Optical Private Network (LHCOPN) and the Large Hadron Collider Open Network Environment, which have core nodes at the StarLight Facility. iCAIR has supported the development of these networks, including core nodes at the StarLight facility, which provides interconnections to national and international networks and a metro area network connecting StarLight to local and national laboratories. Key partners for iCAIR HEP initiatives are the Fermi National Accelerator Laboratory and the Argonne National Laboratory. ICAIR is also participating in current CERN initiatives that include grand challenge data moving planning for the forthcoming High Luminosity LHC (HL-LHC), developing a Network for the Transport of Experimental Data (NOTED) based on AI techniques, and developing a method for optimizing workflows with a packet marking technique.
Astrophysics
With its research partners, iCAIR is participating in several projects addressing infrastructure requirements of astrophysics, especially high-performance networking, including capabilities for supporting specialized research instruments and techniques over international multi-domain research networks. For example, iCAIR collaborates with the international networking research community to prepare for providing communication services for the Square Kilometer Array (SKA), which will have radio telescope sites in Western Australia and South Africa and will generate more data that the LHC, which will have to be distributed worldwide.
iCAIR is also assisting with plans for the Pierre Auger Observatory's Large Synoptic Survey Telescope (LSST), which is being implemented on a mountain in Chile and interconnected by high-capacity international networks. iCAIR also has assisted in developing networking capabilities to support the Sloan Digital Sky Survey, which produces 3D digital astronomical maps.
Another project was an international collaboration established to develop a space geometric technique - very long baseline interferometry (VLBI), which allows for precise measures of the motions of the Earth. VLBI measures the Earth's orientation by placing it within an inertial reference frame. VLBI is based on radio telescopes. By placing antennae in different locations around the globe, collecting radio waves from distant quasars, and measuring differences in arrival times (with picosecond precision), VLBI methods can measure various movements of the Earth. VLBI techniques require the gathering and distribution of large amounts of data.
In addition, iCAIR has also been developing new methods for implementing astrophysical modeling and simulation (on distributed infrastructure based on L1/L2 communication paths on optical fiber) using techniques such as adaptive mesh refinement (AMR).
BioInformatics
With multiple national and international partners, iCAIR has been involved in multiple BioInformatics projects, including one that created a Bioinformatics Software Defined Network Exchange (SDX) or BioSDX, which has been designed, deployed, and demonstrated by a multi-organizational research consortium to enable bioinformatics knowledge discovery supported by dynamic networking services. This BioSDX uses precision networking to support precision medicine.
The BioSDX is based on recent technical developments in infrastructure abstraction that enable new tools and services utilizing programmable network infrastructure through high levels of resource virtualization. Combined with the close integration of programmable cloud computing facilities, the BioSDX prototype is an important advance in supporting the new paradigm of data-intensive bioinformatics across multiple disciplines, including computational genomics and precision medicine—those related to advanced medical imaging and high-performance optical networking.
Also, as a participant in the OptIPuter project, iCAIR developed new techniques for supporting the BioInformatics Research Network project (BIRN), sponsored by the National Institutes of Health (NIH). These techniques allow scientists generating multi-gigabyte data objects at diverse locations to locate, correlate, analyze, and visualize them.
Computational Genomics
iCAIR has also been participating in several projects related to BioInformatics that are developing high-performance computational and communications infrastructure for Structural Genomics, including those supported by the Open Science Data Cloud (OSDC). This multi-petabyte science cloud serves the research community by co-locating a multidisciplinary data commons containing many TB of rapidly increasing scientific data with cloud-based computing, high-performance data transport services, virtual machine images, and shareable snapshots containing common data analysis pipelines and tools.
The OSDC has been designed to provide a long-term persistent facility for scientific data and a platform for data-intensive science, allowing new types of data-intensive algorithms to be developed, tested, implemented, and used over large sets of heterogeneous scientific data.
iCAIR has also supported the development of data transport from the Advanced Photon Source at Argonne National Laboratory to sites worldwide.
Materials Science
In partnership with Northwestern's Materials Science Research Center, iCAIR designed and developed technologies to support new applications in materials science; a joint project funded by the NSF developed the International Virtual Institute for Materials Science, which had all the functionality of a physical research and education institution.
The IVIMS required high-performance capabilities for instantaneously discovering, gathering, integrating, and presenting different sets of resources for a global set of users from around the world. These resources include large-scale data streams from experimental repositories at remote locations, scientific visualizations and digital media, and computational processes.
Nanotechnology
Nanotechnology is the science and technology of precisely controlling the structure of matter at the molecular level. Often regarded as a significant technological frontier, this discipline studies materials and devices at a nanoscale (a nanometer is one billionth of one meter). Several iCAIR initiatives have investigated technologies required by computationally intensive nanotechnology research.
Northwestern University has established the Institute for Nanotechnology as an umbrella organization for large-scale nanotechnology research efforts. The Institute supports major research in nanotechnology, provides state-of-the-art nanomaterials characterization facilities, and fosters individual and group research directed at resolving key problems. The Center for Naofabrication and Molecular Self-Assembly was established as part of this effort.
iCAIR partners with Northwestern's Nanotech Center for Learning and Technology (NCLT) to develop, implement, and operate a large-scale distributed infrastructure to support Nanotechnology science and engineering activities.
Medical Imaging
iCAIR has been engaged in partnerships that explore new mechanisms to use advanced digital media techniques, including imaging, for biomedical applications in cooperation with Northwestern's Feinberg School of Medicine, the National Institutes of Health (NIH), the Radiological Society of North America (RSNA), the Metropolitan Research and Education Network (MREN), national research and education networks, StarLight, and various international networks.
iCAIR and MREN have provided advanced networking capabilities to the annual RSNA conference in Chicago at the Metropolitan Pier and Exposition Authority's McCormick Place, enabling the showcase of new medical imaging techniques. With Northwestern's Feinberg School of Medicine, RSNA, NIH, and the MPEA, iCAIR produced international multicast events on image interpretation. iCAIR also supported professional associations broadcasts of surgical techniques.
BioInformatics
iCAIR has been involved in multiple bioInformatics projects, primarily those related to advanced medical imaging and high-performance optical networking. As one of the partner institutions in the OptIPuter project, iCAIR is developing new techniques for supporting the BioInformatics Research Network project (BIRN), which is sponsored by the National Institutes of Health (NIH).
The OptIPuter project, led by Cal-IT2at UCSD and EVL at UIC, is a five-year, National Science Foundation-funded project interconnecting distributed storage, computing, and visualization resources using photonic networks. These techniques allow scientists in diverse locations to generate multi-gigabyte data objects to locate, correlate, analyze, and visualize them.
Currently, BIRN is a multiscale federated repository for brain imaging. However, the project will be expanded to include other organs.
GeoSciences
iCAIR has participated in several projects on developing advanced optical networking techniques for supporting GeoSciences, which also requires utilizing large-scale, highly distributed 3D objects.
One project for which techniques were developed is the NSF's EarthScope, which involves the acquisition, processing, and scientific interpretation of satellite-derived remote sensing, near-real-time environmental data, and active source data.
Another project measured layers of the Earth's upper atmosphere. A related project developed techniques for oceanography.
Fusion Energy
With its research partners, iCAIR has developed international testbeds to model data flows related to fusion energy research based on tokamaks reactors, including the ITER (Latin for "the way"), the largest experimental fusion reactor currently in operation. A key goal of fusion reactor development is proving that fusion reactions can produce significantly more energy than that provided to initiate the reaction process, producing positive power.
Tokamaks integrate heating processes, powerful magnets, and round reactor containers to spin charged particles and generate extremely hot plasmas to provide energy-releasing fusion reactions in extremely heat-intensive plasmas. Designing and operating tokamaks requires high-capacity, high-performance international networks.
Advanced Digital Media
iCAIR established multiple research and development projects in advanced digital media. Digital media has become an important driver application for the next-generation networking technology design and creation. iCAIR and its research partners have been advancing digital media technology through multiple initiatives that are bringing capabilities supporting high quality, high performance digital media over wide area networks, including internationally.
iCAIR has undertaken projects related to multiple ultra high resolution digital media modalities: 3D scientific visualization; digital-media-on-demand, interactive access to repositories of digital video and related digital objects, which can be directly streamed for immediate viewing or scheduled to be transferred at specified times; Digital media streaming, direct transfer, for live transfer of digital or streaming from archived video allowing for interactivity such as pause, forward, and reverse; digital media conferencing, multi-way interactive high quality video and audio for collaboration among multiple sites, along with supplemental capabilities for additional transmitted materials, such as projected 3D objects; and immersive virtual reality spaces projected over thousands of miles.
In addition, the Center has developed access methods, such as the Digital Video Portal, a research project focused on interactive, network-based digital media. iCAIR also supports networking for the Amart Amplified Group Environment Scalable (SAGE) project at the Electronic Visualization Lab of the University of Illinois Chicago. Also, iCAIR established a research partnership with C-SPAN, enabling its channels to be multicast worldwide. This project has allowed C-SPAN to be multicast at high performance over national and international networks. iCAIR also participated in a consortium that developed an international, flexible, large-scale digital media network based on dedicated lightpath channels within a global optical fiber. For many years, this High-Performance Digital Media Network (HPDMnet) provided for exceptional quality services worldwide.
Ultra Large Science Data
With multiple partners, iCAIR has investigated new methods of using advanced optical networks to support extremely large collections of digital information, especially for science communities. These research projects have examined new lightpath network architecture methods to support extremely high performance data streaming from multiple large-scale data repositories.
Photonic Empowered Applications
iCAIR's advanced networking infrastructure design and development projects are directed at Photonic Empowered Applications—a new class of applications based on extremely high-performance optical communications. "Photonic-Empowered Applications" refers to multiple, global, next-generation applications designed and developed to utilize highly distributed facilities (including those based on resources at sites worldwide). These are resource-intensive applications - e.g., computationally intensive, bandwidth-intensive, storage system-intensive, etc.
However, they are also distinguished by their utilization of advanced data communications based on dynamic multi-wavelength lightpath provisioning and supported by more flexible DWDM-based networking technology than that implemented in today's static point-to-point optical networks. These techniques can transport large amounts of data directly on lightpaths over global fabrics.
They are also optical network "aware" - that is, they can directly discover and signal for the use of the networking resources that they require, including signaling for the provisioning of lightpaths. In addition, some of these applications may be highly periodic and transient (e.g., they may exist only for a few moments at different times throughout a month or a day). Consequently, they may transition instantaneously from a state requiring little or no network utilization to one requiring enormous network resources for days, hours, minutes, moments, or even milliseconds. Many of these applications require close integration of core and edge resources. The boundaries between applications, computers, and networks truly dissolve within this emerging new infrastructure.