Sounds great. It all sounded great until it got to the part about "federal funding". I, for one, have a problem with my tax dollars going towards some professor being able to gawk at another professor in a videoconference.

Part of the reason that this was ok was that the NSF was likely to be picking up the tab for the alternative: a conventional dialed voice call, at about 3 orders of magnitude higher price than the T3 NSFnet's actual costs. People using the NSFnet for AUP compliant voice only saved the taxpayers money. About the broader issue of the Internet II project, I have always suspected that the "retail" market (100 M households) and the R&E market (1000 largest R&E and industrial institutions) had incompatible network requirements. The consequence is that no infrastructure could be optimal for both. It would be a *good*thing* if the providers could develop two parallel infrastructures: one default free, 100k routes, typical peak user data rates near ISDN, to optimally support the retail market. A second parallel infrastructure only needs to support 1000 (rock solid stable) routes, with (local) default routes over to the retail network and sufficient traffic head room to keep the researchers happy. The parallel network would be a premium service, with charges per route as well as per connection. It could be trivially built today (Is being?....) with todays technology, and it could interconnect many of the existing R&E networks (vBNS, NSI, ESnet, etc). Since the R&E networks can scale back the routing problem, they can afford more detailed routing policies and a branchier layer 2, supporting better global path optimization. The hard part is getting all of the peering arrangements in place such that traffic between the premium services of the different providers does not pass through the congested retail infrastructure. Internet II and the like are good things. No question! --MM--