Disclaimer: I've worked with Internet2 for about 8 years. Now I work with the LHC guys, too.
Internet2 has been on Slashdot a number of times. Each time people focus on the network. To me be fair the networking stuff is kind of cool. They're doing some interesting things; tackling some hard problems, providing feedback to hardware vendors that makes their products a bit better, dealing with various political aspects of international networking. All nice things.
However, the networking group is only one of 5 areas within the Internet2 group. Some of the areas are real stinkers and, as best as I've ever been able to tell, produce nothing of value. Others though do. Years ago I saw a presentation from the engineering group about various TCP/IP projects they were working on. Some dealt with intelligent back-off algorithms, some with various aspects of IPv6, a few with QoS that actually worked. The Middleware group, with which I work, has produced some nice work to help educational institutions record basic person information in a standard way (doesn't sound very exciting but it can sure help if you're an app writer). The Shibboleth project, and the related OpenSAML project, deal with making user data available in a secure and privacy preserving way (in theory it wold preclude any more lost laptops with millions of user records). It has seen adoption by various schools, governments, realtor software vendors, and others.
So, the point, I guess, is that it's not just about the network. Another point is that I doubt I2 will be around much longer. The recent failure to merge with NLR, a process that was like watching two of the greasiest used car salemens try to screw each other over, was probably the death toll.
