23-05-2012, 03:24 PM
RESEARCH CHALLENGES FOR THE NEXT GENERATION INTERNET
research_chall.pdf (Size: 345.64 KB / Downloads: 79)
INTRODUCTION
having grown from a research tool intended for a well-defined community to a powerful
communications medium accessed by a worldwide citizenry, the Internet is at
an important transition point. While exciting new applications in medicine, environmental
science, crisis management and other disciplines are tantalizingly near, they
remain out of reach. Today’s Internet cannot scale to meet the number and nature of
demands already placed on it,
much less a new generation of
more complex interactions.
With three decades of experience,
network researchers can
engage in the type of collective
self-reflection that comes from
the ability to draw on the past.
The Next Generation Internet (NGI) initiative is about this transition: using the Internet’s
promise and its past to accelerate the rate of future network development. The Workshop
on Research Directions for the NGI was the first step towards defining a research agenda
for such development.
CONTEXT FOR THE NGI INITIATIVE
Although the research challenges of the late 1990s are significantly different from those of
the late 1960s, one factor has remained constant: the need for government-sponsored network
research and development. Such R&D is as critical to the NGI’s future success as it
was to the success of the original ARPAnet. The ARPAnet’s origins can be expressed in
two words: packet switching. The notion of sending data in packets was a revolutionary
concept developed 30 years ago by researchers at the Defense Advanced Research Projects
Agency (DARPA).
ROLE OF THE FEDERAL GOVERNMENT IN NETWORK RESEARCH
One of the fundamental lessons learned from this series of government-funded partnerships
was an ability to identify those research issues the private sector was not inclined or
willing to undertake on its own. As many workshop groups noted, this particular history
lesson enabled them to identify areas where government-sponsored initiatives could be
helpful to the NGI.
NETWORKING RESEARCH AND THE “SPIRAL DESIGN”
Networking research and development is, of necessity, an iterative process—a “spiral
design” requiring continuous feedback between network researchers and researchers (of
various disciplines) testing applications on the network. These two activities are vertically
coupled: networks drive the applications, and applications drive the networks.
However, the achievement of real progress in network research requires that many applications
be tried. Only in this manner can the common threads—or areas of further network
research—be identified. Of necessity, such testing often takes place using the best
information possible, within a limited period of time, and without the ability to engage in
extended conversation with other researchers.