Internet2 is a consortium framework organisation (a bit like JISC in the UK) within which a large number of projects are cultivated and coordinated. Members are mainly US universities, US government agencies, and significant commercial partners such as IBM and Cisco Systems. Its' purpose is as its' title suggests: to foster the implementation of the "next generation" Internet. A meeting for all members is normally held each spring and autumn.
Internet2 projects divide into 3 main strands:
· Networks - including implementation of IPv6 - the successor to the fundamental communications protocol (the "IP" in "TCP/IP") of the Internet. This will tackle the problem (amongst others) that currently (using IPv4) there is address-space for 'only' about 4 billion directly-connected devices (IPv6 will increase this to 8 x 10^17, or approximately 1,564 IP addresses per square metre of the Earth's surface, which should last us for another few years).
Inevitably (only US institutions can have full member status in Internet2 - JISC/UKERNA is one of a small number of foreign affiliates to the programme) there is a strong bias towards the requirements of the US academic community, and a certain sense that we are picking up the crumbs that fall from their lavishly budgeted and groaning tables; but these are big crumbs! (The network equipment and total bandwidth capacity installed, temporarily, in the hotel meeting venue would turn the IT departments of many moderately sized UK universities green with envy! It's diagrammed at www.internet2.edu/activities/files/spring02networkdiagram.jpg in case you'd like to compare it to your own).
The JISC ANGEL Project (www.angel.ac.uk) is tasked with evaluating the Internet2 middleware project Shibboleth, as one possible model to meet the Sparta requirements for a next generation resource access management service for the UK Further and Higher Education community (www.jisc.ac.uk/pub00/sparta_disc.html). The LSE Library is also being supported by a separate JISC grant to undertake the first beta installation of Shibboleth software, outside of the United States, for trials with pilot groups of users and resources. Significant commercial providers of library resources, including EBSCO and Elsevier, have expressed interest in using Shibboleth and the JISC data service provider EDINA is also likely to be involved in UK beta testing.
As I was supposed to be at the Spring Member meeting to gain familiarity with the Internet2 middleware agenda, and to discuss aspects of the Shibboleth middleware project with the development team, I concentrated my attention on the sessions in that strand - from a programme packed with more than 80 separate sessions (and, of course, not much food and beer at all!) over three days.
The event is partly an opportunity for the widely distributed teams that normally work together via email and voice or video conferencing to meet face to face twice a year. I started the first day by attending an open working group meeting of the Shibboleth team, which turned out to be a poor introduction for 'outsiders', comprising as it did mainly the verbal continuation of a discussion between 5 or 6 'insiders' that had been developing on their email list for the past few months. However, strenuous efforts were made to be inclusive - one of the reasons for opening these meetings is to ensure an element of peer review in the process. And one of the reasons for involving non-US participants in standards development projects like Shibboleth is to try and moderate the emergence of 'global' standards for things like people directories which might otherwise insist that everyone must live at an address with a Zip-code.
There were many more traditional presentation and tutorial sessions that assumed little prior involvement, including "Internet2 101" (which I badly needed!). A session on the CREN initiative to establish an in-house Public Key Infrastructure (PKI) and Certificate Authorities was highly relevant to the current JISC consultation on future authentication for JISC services (www.jisc.ac.uk/pub02/ar1/future_auth.html).
The live demonstrations of development work-in-progress, particularly those involving applications of bandwidth-hungry multiway video were very impressive, but the middleware projects all suffer from the same problem of 'undemonstrability' that we have discovered in ANGEL - users on the end of a Web browser are only aware of middleware when it stops working!
Rather strangely, the 'opening plenary' session was held on the second day of the event, following a host of parallel 'working breakfasts' that provided opportunities for various project teams to meet, and the less-involved of us to discover universities we'd never have suspected the existence of. I had a fascinating conversation over breakfast with a lady from the University of Las Vegas (which, also surprisingly, isn't famous for teaching probability theory).
A session on Directories included announcement of the release for public review of version 1.5 of eduPerson, which defines an LDAP object class to include widely used person attributes in higher education and is closely related to the work ANGEL is doing to define a Roles Namespace for HE and FE in Britain. Current PKI work (both policy and technical aspects) was reviewed. A presentation session on Shibboleth announced the recently agreed participation of EBSCO, Elsevier and a number of other major e-publishers in the project, and also covered related projects like WebISO (Web Initial Sign-On). Ken Klingenstein presented an overview of the NSF Middleware Initiative, with which (it is rumoured) there may be forthcoming JISC calls for collaborative project proposals on offer in the near future.
The closing plenary session focussed on network security issues, with presentations on current work by Howard Schmidt, Vice-Chair of the President's Critical Infrastructure Protection Board, and Mike Roberts, Chair of the joint Internet2 EDUCAUSE Systems Security Task Force. It wasn't surprising to hear that the heightened American preoccupation with domestic security in general is raising awareness of the potential threats of e-terrorism. There was at least a hint of a suggestion that US Government attitudes are shifting away from the previous, "if anyone wants to send a private (i.e. secure) message, they must be doing something wrong, and we must be able to intercept it" of the Carnivore era, and that US politicians feel the need to be better informed about these technologies.
Currently (although it includes digital libraries as one of the applications strand topics) Internet2 seems to be almost exclusively the province of those 'techies' in the networking and IT infrastructure trades. Of over 600 delegates at the Spring meeting, I didn't encounter one who admitted to being a librarian. However, there is clearly an increasing focus of this massive project on middleware, directories and similar issues, all of which could be classified as dealing with metadata about either electronic resources or the people and organisations that use, produce or curate such resources. Luckily, I can still successfully disguise myself as a geek (the older sort, with a beard). But this is the strand of Internet2 that brought me to Arlington, and this is why more people in the library and information management professions need to start taking an active interest in what Internet2 is doing - it will dominate the world!
Details of all Internet2 activities can of course be found at the 'obvious' URL of: www.internet2.edu
John Paschoud is an InfoSystems Engineer working in the Library of the London School of Economics & Political Science, and is the Project Manager of ANGEL.