This is an archived version of CRA's website. This archive is available to provide historical content.

Please visit http://www.cra.org for the latest information.

Computing Research News

March 2006     Vol. 18/No. 2

Back

GENI and Your Research

By Peter A. Freeman Assistant Director of NSF for CISE

 

GENI (Global Environment for Networking Innovations) is an advanced experimental infrastructure and accompanying research program being planned by CISE to explore new capabilities that will advance innovations in many areas.1

GENI responds to an urgent and important challenge of the 21st century to ensure that the future Internet will be worthy of our trust, able to continue to grow robustly, and capable of supporting even more innovation in all areas of activity than the current Internet has enabled.2 Ultimately, achieving this goal will depend on a large number of factors—including legal, regulatory, policy, commercial and technical—but it begins with exploring new networking and distributed system architectures that can respond to the demands of the future.

Nonetheless, this effort will ultimately touch many areas of computer science and engineering (CS&E) research. Because GENI has the potential to touch your research and perhaps change it fundamentally in the long run, and because it will be a major undertaking for CISE, I want you to be fully aware of what is being done and of the potential for you and the field.

We intend that it will provide a platform for innovative research in a number of CS&E fields beyond networking and distributed systems, including databases, operating systems, languages, interfaces, CS and communications theory, robotics, ubiquitous computing, sensor networks, and so on. While we at NSF can’t spell out exactly what those opportunities will be—that is your responsibility—we can outline the characteristics of a project such as GENI that from previous experience we are confident will create these opportunities.

First, as you will see if you look at the GENI conceptual design posted on www.geni.net, the GENI facility is itself a complex system. While the conceptual development to date indicates that the GENI facility can be realized without major research activity, designing and constructing it over the next seven or eight years will undoubtedly uncover compelling research topics in a number of areas.

Second, as GENI is used to experiment with new networking and distributed system architectures at scale, new services and applications will follow very rapidly. These, in turn, will require or uncover new fundamental developments in CS&E, and enable new generations of research in CS&E fields like robotics and artificial intelligence that may not be directly involved with the underlying systems research.

Third, and most importantly in my opinion, GENI will enable and encourage a return to large-scale experimentation—a research modality that was dominant and highly productive in the early years of our field but has waned in recent years.3 Initially, this will be experimentation in networking and closely related fields, but the nature of the GENI facility is such that it will also permit experimentation in other areas, such as distributed databases, as well.

Beyond the instrumental use of the GENI facility, we believe this project will demonstrate the importance and effectiveness of large-scale experimentation in CS&E in general. This will then pave the way for experimentation (and the infrastructure necessary to support it) in other areas such as computer and system architecture.

Even at this early stage we are already seeing another laudable and highly valuable development—the involvement of theoreticians along with the experimentalists.4 CS&E has never really had the virtuous cycle of observation-theorizing-experimentation followed by more observing, theorizing, and experimentation that other sciences have had. We believe the field has matured to the point where it must employ this modality vigorously; GENI and, ultimately, other efforts, will afford this modality.

Another and more mundane issue regarding GENI and your research is that of funding. Constructing the GENI facility will be expensive, and the first question that you might ask is "Why not spend that money on more research like that which is already being supported—small grants." The first and most important thing you need to understand is that money appropriated for GENI construction, if approved, may not be used to support research grants. And CISE funds currently supporting research grants will not be redirected to support GENI construction. While CISE certainly plans to devote future networking research budgets to research conducted using the GENI facility, other programs will not be impacted unless they determine that it is in the interest of their fields to leverage GENI capabilities.

A second important thing to understand is that the nature of science funding—indeed of funding for projects in general—is that it is the ambitious, paradigm-shifting projects that capture the attention of those responsible for appropriating funds. The result is often a general infusion of money into a field that has many positive benefits that are ancillary to the original project.

This country is facing serious challenges in the area of innovation and economic development. CS&E has clearly shown its relevance to innovation (and thus to economic competitiveness) in many ways, but much of the current innovation occurs not in the research community but in the commercial arena. While this is clearly the desired end result, those innovations are invariably based on fundamental ideas and inventions that were developed—usually many years ago—in our labs. We must re-energize the type of activity that a generation ago produced most of the fundamental concepts that fuel today’s commercial innovations.

We will be asking the broad CS&E research community, not just networking and distributed systems, to form a community consortium to guide the scientific and administrative development of the GENI facility and then its usage. Over time, this representative proxy for the CS&E research community will have the opportunity to propose other major projects to support our research. I encourage you to cooperatively support the formation and operation of this consortium. It will enable us to speak clearly and effectively as a community about what we can do and what it will take to move forward.

GENI is the first in what I hope will be a series of major efforts to rekindle fundamental invention in our field. Our nation’s innovative posture depends on it.

I look forward to hearing from you. Please send general comments to me at pfreeman@nsf.gov, or comments specific to GENI to geni-info@nsf.gov.

Notes:

1 See www.cra.org/CRN/articles/nov05/freeman.html for a high-level description. Consult www.geni.net for current and more complete technical information. For NSF plans and announcements, see www.nsf.gov/cise/geni/.

2 Two recent and influential reports (National Innovation Initiative Interim Report www.compete.org/nii/ and Rising Above the Gathering Stormwww.books.nap.edu/catalog/11463.html) have stressed the essential role of innovation to our nation, and the absolutely necessary role of research and education. It is becoming widely understood that the Internet has already enabled a surge of innovation, but it is clear to those most knowledgeable that this cannot continue without some fundamentally new research. This effort, then, is intended to address this situation and to help enable the innovation being called for nationally.

3 The reasons for this are varied and need not be debated here. A return, however, to a modality in which the field includes, but not exclusively, a strong experimental activity is generally considered to be essential for the vitality of computer science and engineering.

4 See ToNC and www.comsoc.org/livepubs/surveys/public/2005/dec/index.html (“News from NSF”), for example.


Peter A. Freeman is Assistant Director of NSF for CISE.

Back

CCC Logo

CERP Logo CRA-E Logo CRA-W Logo

1828 L STREET, NW SUITE 800, WASHINGTON, DC 20036 | P: 202-234-2111 | F: 202-667-1066