Internet traffic growth myths (cont.)

costs
tech stock bubble?
really takes new technologies a decade to penetrate
web was exception, Internet is not
retarded technical developments
negligence of what users what, and likely to get
community gets mired in sub-necessary QOS hubbub, ATM, GMPLS

benefits 
unparalled platform for innovation
open standards, rapid development of new services
big empty pipes was key factor in supporting revolution

lessons
25 year contracts for pipes should be amortized over 3
capacity planning impossible 
simplify engineering (atm/sonet --> IP over WDM, GigE)