Event ended Lecture
Spring 2003 Internet 2 Member Meeting

About this event

[Introduction by Doug van Houweling, President of Internet2]

Thank you, Doug.

It is truly a pleasure to be back here with so many friends and colleagues. It was not quite a year ago - at your Spring meeting last May - that I had the privilege of addressing you on my second day at NSF.

I must say that the intervening ten years has been exciting. It has been packed full of opportunities for me to really screw up!

I think that I've managed to not do that - at least so far - and perhaps have even been able to make a small, positive impact. Of course, we'll have to wait a few years to really pass judgment on that - by which time I'll be gone!

On a more serious note, let me say how pleased I am that so many of you are from countries other than the United States. Networks have a global reach and their development has always involved international collaborations.

It is a great privilege to work at NSF at this very important time in the history of our country. I have had the opportunity to work in a number of roles in academia at several institutions - including as a CIO overseeing a major campus network upgrade. I have also worked extensively with industry, again in a variety of ways.

To now be able to serve my country in a direct way on the national stage is very, very rewarding, and I encourage you to do the same at some time in your career.

It is also a very challenging time in the history of computing in general, and of networking in particular. We have proven the value and transformative power of our technology, but we must now insure that we don't stagnate on a plateau, resting on our laurels.

An aspect of that is what I want to talk with you about this afternoon.

When Laurie Burns asked me several months ago for a title and abstract for my talk, I gave her what you have seen posted. I will say a bit about challenges and what we are doing to help meet them as the schedule advertised, but I really want to spend my time on a broader and, ultimately, more important, topic - community, and the cooperation and collaboration that are the hallmarks of a community.

So that I am not charged with false advertising, let me first dispense with the obligatory - but sincere - comments about our NSF programs.

Let me also note that my colleague, Dave Nelson, Director of the National Coordination Office for Networking and IT R&D, in his closing plenary on Friday will describe not only some of the NSF supported activities in the networking area, but those of a number of other agencies as well; I encourage you to attend his talk.

I said to you last spring that networking has been, is, and will continue to be one of the most important topics for NSF and CISE. In the past year we have

  • Increased the number of competitive awards in ANIR for FY02 to 160. This is more than a 30% increase over 2001.
  • Total grant funding for ANIR in FY02 was $81.5M
  • NMI had 11 new awards for about $2.5 million
  • The latest cooperative agreement for ETF included $14M for networking
  • $10M is projected for FY03 ETF extensions program, which is largely networking
  • 2 new ANIR programs were announced (Network Research Testbeds & Experimental Infrastructure Networks)
The EIN or Experimental Infrastructure Networks Program will support projects that explore innovative uses of existing network technology to enable an application that has performance or other requirements that cannot be satisfied on a shared IP network.

The NRT or Network Research Testbed Program will support the implementation of testbeds that explore exciting new approaches to building or understanding tomorrow's networks.

The Extensible Terascale Facility (ETF) Program, supports the development of an integrated, grid-based environment that will included cutting-edge computational, storage, sensing and experimental resources. A very high performance network is its foundation.

Cybersecurity is another topic of great importance to all of us, and one on which you probably spend significant time. If you don't, you should! Cyber security will challenge us - "us" broadly, from networking to systems to information security workers - to take ideas from research to development and on to effective deployment, working as a community on this pressing problem.

The Cybersecurity R&D Act passed late last year authorizes NSF to take the lead in research, education, and building capacity for education in this area. While new appropriations of the magnitude envisioned have not yet been made, we are moving as much new money into this activity this year as possible and will soon be announcing several activities to move us forward even more quickly. The challenges for the networking community are clear to us as they are to you. I will come back to this in a few minutes.

On the cyberinfrastructure front, we received the Atkins Report in late January. It articulates a vision that, in fact, is the natural evolution of the computing, networking, and software investments that NSF has made for years. We are continuing to move forward to make that vision a reality as rapidly as we can.

Cyberinfrastructure isn't something new, except in its extent, capability, and interconnectedness - which is exactly where networking comes in. Networking is absolutely the foundation on which CI is being built. So, your part in CI is assured - whether you like it or not!

In case you think that our current stock of ideas and technology is sufficient to build CI, I suggest you look at some of the large-scale S&E projects underway or being planned by NSF and other agencies - NEES, NEON, Earthscope, the Large Hadron Collider - and ask yourself if we can handle their network demands? We have significant research and operational challenges ahead of us that will take a lot of inspiration as well as perspiration. I believe that Dave Nelson will describe some of these in more detail on Friday.

This is a good time to note just how international science is. For example, projects based on the Large Hadron Collider at CERN will include scientists from many countries. They will use cyberinfrastructure that has been made possible by international collaborations in fields such as middleware, distributed computing, networking, database systems, computer architecture and others. To build, operate and effectively utilize the future cyberinfrastructure will require continuing collaboration between scientists, computer scientists and operational personnel throughout the world.

There is certainly no end of networking challenges - CI will require networks that allow scientific collaborators to discover, access and share resources on an unprecedented scale and allow geographically distributed groups to effectively work together. These next generation networks will exhibit unpredictable and complex behavior and dynamics. Advances must be made to create and sustain the science and technology needed for the effective engineering, control and management of a ubiquitous, scalable, reliable, adaptable and secure network infrastructure.

We will need continued fundamental research and experimental development, building cyberinfrastructure, and making sure that we can secure it all. We are working hard at NSF to focus our activities in networking and to obtain the increased funding that is certainly needed.

Now, I hope I've fulfilled the expectations for the advertised talk, at least minimally. So, I could just quit now and you could all go catch up on your email - or enjoy a stroll outside on this beautiful(?) Spring day. On second thought, given the weather we are experiencing here in Washington, you may want to just stay here!

I want to spend my remaining few minutes talking with you about an even more important - and very much related - topic: Community, and the cooperation and collaboration that goes with a community.

Let's perform a small gedanken-experiment together. Imagine viewing a network in a zoom-out mode such as we see on TV accounts of the war.

From the perspective of a user, a network might be seen as a system of computers and other devices interconnected by lines and routers to provide shared access to data and resources. Even this simplistic (compared to your sophisticated understanding ) view of a network is more detailed than it might be.

Zooming out further, a network can also be seen as an extended and dispersed group of people with similar interests or concerns who interact and remain in informal contact for mutual assistance or support. As I considered how to address the networking challenges that we all face, I kept returning to this very general kind of network, that folks like us so seldom consider.

And then it struck me; it's not the technology. It's the COMMUNITY, stupid!

My fundamental message this afternoon, then, is very simple: Networking is about much more than transporting bits; it is about building communities.

I think - I hope - that many of you already recognize that just moving bits around the world, or around your campus, is a pretty sterile objective that in the bigger picture, hardly anyone cares about.

What they care about is what it permits them to do: Communicate with others instantly and widely, control and utilize scientific instruments at great distances, access data in volumes and to an extent that, literally, no human has ever done, work collaboratively on projects whether across campus or across town or across the world. Many of you come into contact daily with the scientists and engineers on our campuses - to say nothing of those in other, less technical disciplines - who are doing this. While that must certainly give you a birds-eye view of what they are doing with your networks, I want you to step back and reflect with me a bit on the underlying meaning of what they are doing.

Let's first look at some dictionary definitions that I believe will provide further clues to what is ultimately important about networking.

If you look at the definition of "community" in the dictionary, it focuses on the physical proximity aspect of community that is the common understanding of the word. If you ask someone on the street "What community do you belong to?" they will probably answer with something that is either strictly geographic - "I live in the Greenwood community" - or something that is pretty closely delimited by geography - "I belong to the Washington medical community."

One of the oldest secular communities that was not bound by geography is the academic community - enabled by shared values and the ethos of communication among scholars not only locally but also globally. Similarly, there have been religious communities for even longer - witness the broad reaches of the Roman Catholic Church and, in a less tightly organized but perhaps even more cohesive manner, the world-wide Jewish community.

As communications have improved over the past century - at an exponential rate of increase - secular communities have formed that are not as restricted by geography. Thus, we find people talking about belonging to the medical community or the Irish-American community or the liberal community. These secular examples have evolved, at least in concept, from the older idea of a widespread, academic community in which frequent communication enables achieving objectives that could not be achieved by the individual. Obviously, networking, especially in the past 10-15 years, has permitted many communities to form that never existed before - because of the ease of communication.

But, communities are not just groups of people that talk a lot. They are groups of people who have developed shared values and goals through their communications.

This, in turn, leads to cooperation - working together toward a common goal that typically cannot be attained by an individual or a small group of individuals, as the dictionary tells us. But this cooperation requires more communication, which is facilitated by our modern means of communication, which can lead to a yet stronger sense of community in a virtuous spiral.

Another term that is in frequent use today is "collaboration." The dictionary indicates that collaboration is a form of cooperation, especially on intellectual tasks - a meaning we are all familiar with because of it's common usage in academia when we speak of our fellow scientific "collaborators." And, of course, the term "collaboratory" is in fairly common usage to indicate a physical laboratory that can be used remotely and/or in an arrangement that permits scientists to collaborate remotely - both instances of cooperation that are enabled by you and your networks.

(The dictionary also notes the connotation of "collaboration" used in wartime when one speaks of "an enemy collaborator." But, we won't go there!)

My purpose today is not to give you a vocabulary lesson, but to set the stage to discuss with you the importance of community and cooperation, and to suggest some ways in which you can have more impact. To do that, let's focus on the science and engineering research community - a community that you work with on a daily basis and that is the focus of NSF's activities.

As the lawyers would say, let's start by stipulating that community and cooperation are good things in science. Physicists couldn't perform their gigantic experiments without cooperation. Biologists couldn't sequence the human genome without cooperation. Computer scientists couldn't build complex systems without cooperation. You can't build and make secure campus or metropolitan networks without cooperation. And on and on, ad nauseam.

We aren't going to be able to build the integrated cyberinfrastructure that is essential to the progress of all areas of S&E without cooperation between the computational scientists, computer scientists, network engineers, data curators, etc. While CI is seen by many as an enabler for communities, it can only be created in the first place by prodigious cooperation. Each of you will have an opportunity to help with that.

As one example, it seems clear that many campuses have not solved the "last mile problem." More than a few times in the year I have been at NSF I have had faculty tell me that while their campus may be well-tied into high-performance networks, that that doesn't make much difference to them because they can't really get much bandwidth to their labs and offices.

Are you regularly talking with the S&E researchers on your campuses? I don't mean just making sure that your help desk is well-staffed. I mean really being in touch with those users, getting inside their heads and understanding what their attempts to use your networks are really like and what they know they will need soon. I mean communicating not just with the hand full of high-end users, but with all kinds of users of your networks. Just because high-end users are well-served, doesn't mean that everyone is well-served.

Closer to home, you can't build and operate the networks that will be needed for CI without cooperation. I'm sure you have a lot of cooperation going on now, but let me ask:

Are you cooperating with the network researchers on your campus and/or in the larger community - there's that word again! - to understand and try out the latest ideas and in the process to help develop those ideas? My guess is that the answer is "no," or "not very much."

Before you start reciting the reasons why that won't work or how you've tried and failed or how you just don't have the time or how they aren't interested, let me note that I have probably heard all of those reasons. Remember: I have been a faculty member and a CIO, as well as a dean. I understand from both sides the mutual disregard that all too often exists between faculty and the computing facilities organizations on a campus.

I understand that there are reasons that both sides contribute to why this communication is sometimes less than successful. No one ever said that communication and cooperation were easy. But, I know that without increased cooperation between the networking research community and the operational community we will continue to fall short of the needs and expectations of your users and our efforts to develop campus, regional and national CI in support of e-science will be less effective.

Let's look at a third area that cries out for more cooperation: cybersecurity. Are you working to get the cooperation that is needed between the technical, administrative, and academic communities on your campus to address security? Are you working with the security researchers and educators on your campus? If there aren't any, are you working with your CS&E department(s) to encourage some faculty to start working in this area? Are you exploring with them the possibility of and mutual benefit to both of you of using the campus network for properly controlled experiments and as a training ground for security professionals?

As a last remark on this theme, let me note that the new programs I have mentioned - Experimental Infrastructure, Network Research Testbeds and Extensible Terascale Facility programs - all will be focused on projects that enable new communities, that require deep cooperation and nourish new opportunities for collaboration.

By this point, I hope that you see my point:

You are helping to create communities that encourage and enable cooperation and collaboration. This is essential to the advancement of S&E. But, that cooperation must also be applied to what you do. Find ways to cooperate with the groups on your campuses that can contribute to or inform your work. Together you will be able to do things that neither of you can do alone.

This continues to be an exciting time for our field, but as in so many activities in this world, we must learn to cooperate and collaborate more effectively if we want to achieve the objectives we can envision.

I look forward to hearing your success stories the next time we meet.

Thank you.