1.3: Network Infrastructure
Evolving and Scaling Socio-Technical Network Infrastructure Other Information:
Network connectivity, beginning at the computing platform or device and radiating outward through local-area networks connecting
to wide-area networks connecting to the Internet’s network of networks, creates the wired and wireless communications fabric
that makes a digital world possible. Over the last four decades, Federal investments in basic network research have led the
way to the Internet, the World Wide Web, wireless mobile and optical networking, and an array of network-based applications
that are reshaping societies and economies around the globe. Where we are now: Advanced computer networks provide the infrastructure
for transporting, developing, archiving, accessing, and using the huge volumes of data that support critical functions in
every sector, such as storage and nearly instantaneous interchange of data in the financial markets. Scientific experiments
such as the Large Hadron Collider at CERN distribute petabytes of data to thousands of scientists around the world who are
seeking to uncover the fundamental nature of matter and the “dark energy” that dominates the universe. Dynamic, heterogeneous,
secure, and reliable networks are also critical to DoD’s ability to defeat adversaries, to DHS’s ability to respond to natural
disasters and terrorism, and to Federal efforts to improve health care for all. Like the LHC, a growing number of scientific
applications require very large bandwidths to support massive data transfers and the need for near-real-time coordination
and data transmission protocols tailored to the data requirements. Other such applications include Earth system modeling supported
by the Earth Systems Grid (ESG); computational genomics; and Very Long Baseline Interferometry (VLBI), a radio astronomy application
enabling simultaneous observations of an object by many telescopes combined, emulating a telescope with a size equal to the
maximum separation between the telescopes. Current high-performance networking (including science networks) does not generally
support these demanding requirements; the current approach is to provide dedicated point-to-point network links and services
among the key researchers (or to have researchers move to sites where adequate networking is available). Today, such services
lie mostly outside the science networking provided to the larger research community and are implemented outside the university
science networking infrastructure. Research needs: The network infrastructure of the 21st century must be made robust enough
to meet very diverse demands, including network services supporting A7; exponential increases in data volumes and changes
in how people access data (e.g., data in the network); ultra-reliable, secure networking (e.g., for national security and
the commercial and banking sectors); new networking technologies that scale (e.g., all-optical networking) to provide the
end-to-end bandwidth, performance, and services required for data-intensive science; and heterogeneous networking (e.g., wireless,
optical networking, satellite communications, and others). The following are core areas of networking R&D in which advances
are needed: * Foundations: Architectural principles, frameworks, and network models to deal with complexity; heterogeneity;
multi-domain federation, management, and transparency; end-toend performance; and differentiated services. * Design: Secure,
near-real-time, flexible, adaptive services with built-in intelligence to facilitate discovery, federation, and management
of resources across domains and to increase the application robustness and resistance to attack even in extraordinarily complex
systems and new ways of interconnecting networks to provide those services. * Management: Management methods and tools that
incorporate intelligence in the network to enable effective deployment, control, and utilization of complex networks and resources
in dynamic environments, across domains, and with ever-increasing network and application complexity. * Privacy and Security:
Achievement of a high degree of security even in complex, heterogeneous federation and policy environments, especially in
the face of component failures, untrusted components, malicious activities, and attacks, while also respecting privacy and
maintaining usability, e.g., provide scalable federated policies for authentication, authorization and accountancy * Usability:
Adaptable, user-centered services and interfaces that promote efficiency, effectiveness, and fulfillment of user needs without
overwhelming users with unnecessary or unauthorized data. This agenda must be pursued across the spectrum from fundamental
to applied research and with engagement of all sectors to attain widely deployable innovations. Necessary elements include:
* Basic and applied research in the full range of network architectures, theoretical models, analysis techniques, hardware,
software, security and privacy, and middleware needed to support the next generation of uses for networks and explore new
paths to develop capabilities that cannot be supported on the current evolutionary path * Partnerships with application developers,
users, and stakeholders to test basic research ideas on real problems in areas including national security, support of scientific
leadership, and human health * A suite of testbeds and prototype networks that enable understanding and creation of new technologies,
data systems, and improvements in end-to-end performance at varying scales. The massive size of existing deployed networks
such as the Internet limits research and development, while laboratory and simulation studies cannot address some aspects
of the solutions, particularly complexity, their ability to scale, and their potential realism. The testbeds and prototypes
will range from high-flexibility/low-cost platforms to highperformance embedded systems. Research, partnerships, and testbeds
and prototype networks are closely interrelated. Testbeds and prototypes are needed to test and develop new networking capabilities
in realistic environments, to assure they can be implemented technically and economically, and to explore policy frameworks.
Partnerships between the researchers and the application developers will help assure that R&D capabilities can be implemented
in real networks and that other application resources such as computing and storage are provided.
Indicator(s):
|
