The technology, of course, has its strongest roots in research and academia -- it began as a joint development effort between the Argonne National Laboratory, the University of Chicago and the Information Sciences Institute at USC, and continued with the partnership of similar institutions in the Globus Alliance. Today, many vendors who actually contributed to these big science and research Grid implementations are now interested to usher the Globus Toolkit into enterprise Grid scenarios as well.

Having personally stood on both sides of the bridge, first at Argonne National Laboratory and now with the enterprise vendors in the Globus Consortium -- I've observed firsthand the challenges that a technology developed under the auspices of educational and government grants faces in the transition to industry adoption. These truly are two different worlds. We often have similar IT challenges and requirements, but the pace, culture and motivations are remarkably different.

It's unfortunate that the academic / research and enterprise Grid evolutions are often portrayed as being mutually exclusive somehow. Enterprise reporters and analysts tend to look wearily at research / academic grid efforts and scoff that they're 'just science experiments' -- not applicable to mission critical, enterprise settings. Academic / research audiences sometimes cast a weary eye towards the enterprise, and wonder if the commercialization of the Globus Toolkit would somehow bastardize what has to date been a very open, collaborative free software development movement. Both of these outlooks are unwarranted and counter-productive, but they do exist on some level.

We've seen this sort of sibling rivalry between the two realms in the early history of other major technology shifts. The criticism of the Internet in the early days, for example, was that it was that it was "just a science experiment." It turned out that those science experiments laid the foundation for what turned out to be a pretty great thing for both research / academia and the commercial world. We hope that the Globus Toolkit can play a similar function for Grid adoption that TCP/IP played in the advancement of the Internet. The Globus Consortium members certainly recognize and appreciate the exhaustive Globus Toolkit development work that's come from the Globus Alliance and the Department of Energy.

This month's issue of the Globus Consortium Journal includes some commentary (in Ian Foster's column, as well as in the Q&A with Eben Moglen) about open source licensing issues, and why Globus Toolkit recently embraced the Apache 2.0 license. At face value, this may appear to be legal mumbo jumbo -- but it's truly compelling in the sense that all Globus Toolkit decisions today must now account for the best interests of such a wide range of constituents. Scientists, academics, enterprise vendors, enterprise end users, open source developers, government ... the universe of technologists who have made a commitment to the Globus Toolkit at their respective organizations has become considerably larger over the last ten years.

The open source nature of the Globus Toolkit is the foundation upon which the bridge between all of these different audiences and interests was built. As the Globus Toolkit gains wider adoption we will see a future where these interests come together on a more frequent basis to work towards common goals.

-- Greg Nawrocki

close window