The Grid as a working horseAmsterdam 11 October 2002 In Primeur/EnterTheGrid, when reporting on the Grid developments in the USA, we usually report on the bleeding-edge projects like TeraGrid and the alliances administered in Urbana-Champaign and San Diego, that are involved in Grid and supercomputer research. However, most scientists in the USA, as in Europe and Asia, are not interested in being involved in Grid or supercomputer research. They want to do their bit of science and excell there. When Grid computing helps, fine, if not, they use another tool. In most countries, there are user organisations of collaborating universities that try to help scientists with the infrastructure they need. In The Netherlands that is SURF (no, not SURFnet, I will explain that later) and in the US that is Internet2. Internet2 and SURfnet were involved in the organisation of iGrid2002 in Amsterdam.
In the Netherlands, SURF is the organisation of collaborating universities and academic institutions that is involved in discussions and planning on infrastructure and other IT matters.
Maintaining and developing the actual research network itself is spinned-off into a separate organisation called SURFnet. (See also the separate article on SURFnet in this issue).
In the USA, the user organisation is called Internet2. This organisation is a "user group" that has over 200 members, mostly universities and a few research institutes. There are two differences with, for instance, the Dutch situation. Internet2 is financed by its members, not by the government, and Internet2 develops and maintains the network for US researchers.
As Heather Boyles, Director of International Relations of Internet2 explained, the main goals of Internet2 are to put a good network environment in place for researchers and to develop new Internet applications.
Most university campuses have a good infrastructure in place. Internet2 connects these campuses. On the lower network levels this is already done in a transparant way. When it comes to really using your campus resources transparently off-campus, when you are not physically there, is still a problem. Scientific publishers, for instance, use IP-numbers to check whether you are allowed to viewtheir publications on-line. "So", Heather Boyles explained, "being here in Amsterdam, I cannot view the journals I am entitled to contractually, because I am not on the campus." It are these types of problems that hinder collaborations between researchers, and deprive researchers from resources when they are not at their university.
Hence, authentication/authorisation is one if the important items on which Internet2 focuses in the application area. They are involved in projects like Shibboleth that tackle these problems.
The idea is that all campuses already have their own authentication method. The goal is to make them interoperable, not to impose a new standard.
Internet2 also showcases advanced technologies to show to its members what can be done. The iGrid200 demos with for instance use of iPv6, HDTV over the ocean, and remote use of a telescope in Osaka, Japan, are useful examples.
In the USA, they are looking, of course, to the tools developed by the NPACI and Alliance supercomputing, and Grid tools and applications. The NMI in the USA is developing usable core middleware, and has a Grid support centre in place.
Ad Emmen [News on Advanced IT] [Calendar] [Analysis] [IT in Medicine] |
 |