Yesterday’s announcement that ESNet is buying into the National LambdaRail as its backbone provider of the future is excellent news, but not exactly a surprise. More than ever, national laboratories and academic research communities need to collaborate, and making sure that they share a common network fabric in the future will certainly help facilitate that. This hookup has been coming for a while, though. Suggestions in this direction have been showing up in DOE slideshows since 2002 at least. The DOE Network Roadmap (late 2003) notes that “As NLR evolves, it will be important for ESnet to interface with the services running on NLR, as many of the users of Office of Science facilities are located at universities, and these users will use these high end services as their method of moving terabyte/petabyte-scale volumes of data to and from ESnet.” Bill Johnston’s reviews of ESNet in early 2004 (e.g. here) make connecting ESNet and NLR a critical near term goal. As usual, big data sets (e.g. LHC experiments at CERN) pull this train.

InformationWeek’s Aaron Ricadela looks at the “seismic shift” occurring in the world of high-end computing, pointing to the retirement of the PACI program and the National Science Foundation’s new focus on “shared cyberinfrastructure.” The article examines the unique roles each HPC center will play, with San Diego focusing on data, NCSA focusing on innovative systems, and Pittsburgh supplying cycles.

In a sidebar interview, Ricadela talks to Peter Freeman, leader of NSF’s Computer and Information Science and Engineering Directorate. Freeman describes the shift taking place at NSF and how shared cyberinfrastructure will benefit users. “When people talk about PACI, they’re talking about the way high–performance computing used to be provided to the research community, versus the way we think it will be provided five to 10 years down the road,” Freeman explains. “We’re right in the middle of that kind of transformation.”

A second sidebar looks at NCSA’s plans for the future under new director Thom Dunning, including the development of “cyberenvironments” and workflow tools to assist researchers, an expansion of the center’s Private Sector Program, and the launch of the Innovative Systems Lab to push the computing envelope.

As the title of the NSF program makes clear, the research community thinks of cyberinfrastructure as something people essentially share. An article from last month’s Economist about the increased interest in “sharing” as an economic phenomenon seems to be part of the same movement, especially since it’s been stimulated by things that are directly related to cyberinfrastructure – the open source software movement and peer-to-peer technologies. As you’d expect given their audience, the Slashdot posting on this, which linked to Yale’s Yochai Benkler, a leading thinker in this area, and his newest paper on the subject, drew a significant number of responses.

I think one way to separate out the research community’s conception of cyberinfrastructure from the vendor’s conception, on which we posted earlier, is to ask “Who is sharing what with whom, and for what purpose?”

NetworkWorld has published the transcript of a grid computing panel discussion among leaders from various commercial vendors, including IBM and Cisco. While the discussion has an obvious commercial slant, some interesting thoughts and opinions about the pros and cons of grid computing are shared by some of the leading players. Some of the responses to the moderator questions truly expose the position of some vendors, which may or may not be common knowledge.

A research fellow at Fujitsu, speaking in InfoWorld, says that the company is the first chip manufacturer to commit to using carbon in place of copper. Nanotubes carry mitigate electrons’ tendency to leak out of copper wires, transmit electrons 10 times faster, and dissipate heat more readily, according to Fujitsu’s Yuji Awano.

More on the company’s nanotech strategy from the house magazine.