Andrew K Bjerring and Bill St Arnaud

The Internet is Transforming How We Conduct Research, Part I

By Andrew K Bjerring and Bill St Arnaud

In 1985, the National Science Foundation (NSF) created the world's first Internet backbone to link its five supercomputer centers to each other and to their respective user communities at universities across the United States. Virtually overnight, researchers began to communicate with their peers around the globe faster, more frequently and more completely than had ever been possible with scientific societies, conferences and peer-reviewed journals.

The past 15 years has been the first "Internet generation" for research.

This year, the same NSF is supporting several new "grid" initiatives to install and network thousands of live sensors in ecologically-sensitive forests and ocean floors to permit on-going monitoring and analysis by a research community that will effectively span the globe. Related projects in several countries are seeking to link hundreds of computers in "computing grids" that will surpass in computational power the most sophisticated high performance computers ever created. Others will seek to use new data distribution standards and techniques to provide access to hundreds of terabytes of high-energy physics, biological or earth observation data. Still others will use the reach and power of the Internet to distribute inter-related elements of a single, highly parallelizable research problem to tens of thousands of participants.

These initiatives constitute a clear advance on the past 15 years of enhanced collaboration, and constitute a second Internet generation for research. Each of these grid projects is, in a distinct way, empowering researchers to actually couple their respective research programs more tightly with others. The Manhattan Project of World War II may be the best example of a successful assault on such a "grand challenge" problem. But if these initial grid projects on the part of NSF and other agencies prove successful, grand challenge research may become more common in fields beyond high-energy physics and engineering, and could have implications for education.

Underpinning these second-generation grid projects is the phenomenal increase in the power of the Internet and the increased power and sophistication of workstations. For the past seven years, CANARIE has been working on the networking dimension, including building the world-leading CA*net 3 backbone network and its role in building 10 provincial counterparts. Nevertheless, there are still regrettable gaps in the nation's research intranet that leave individual researchers with on-ramps of inadequate capacity. Closing those gaps will continue to be a challenge for campus planners and the Canada Foundation for Innovation, as well as CANARIE.

Such gaps notwithstanding, we believe that our researchers are generally well positioned to take part in the various grid initiatives that are being planned. We also realize that the bar is constantly being raised. For the next generation research network, we are planning a network that will allow individual researchers to reserve dedicated wavelengths of multi-gigabit-per-second capacity from anywhere and to anywhere. A switched-wavelength network of this sort, coupled with equivalent projects elsewhere, will enable Canadian researchers to participate in grid projects the world over.

Impressive though the changes in the capacity of networks and workstations have been over the past 15 years, in some respects the real breakthroughs with the Internet are being revealed in other domains. For example, the MP3-based applications, Napster and Gnutella, are remarkable not because of the brute force of the networks that enable them, but because of the way in which they exploit so elegantly the distributed nature of intelligent devices in the Internet world. Analogously, the real power of future research grids may derive from breakthroughs in the way in which distribution is effected. Although such a third Internet generation of research may be speculative at this point, one thing is clear - in research as in so many other areas, the Internet has, once again, changed everything.

Dr Andrew Bjerring is president and CEO of CANARIE Inc. Bill St Arnaud is its senior director, advanced networks.

Part II of the Internet's impact on research will appear in the next issue.