On Thu, Oct 30, 2008 at 6:28 PM, David Meyer <[EMAIL PROTECTED]> wrote: >> Seriously, what passes for cost analysis in this industry is shameful. >> Mostly it's "experts" sitting in a circle making SWAGs. If this time >> is an exception, I'd like to read the paper. > > What I know is public is the following: > > (i). A+P > > http://www.1-4-5.net/~dmm/a+p.pdf
Hi David, So you shift a few bits of the UDP and TCP port numbers into the addressing space, lets say 4 bits, giving a 16x multiplier on the usable space in exchange for each host "only" being able to have 4000 active ports. Anyone with an A+P compliant host gets a global address that preserves the end-to-end principle, at least for UDP, TCP and the related ICMP errors. Legacy hosts get rfc 1918 space, fall into the CGN and still work fine for most stuff. Other IP protocols fail, of course, but then they fail with a CGN too. Very clever. The only word about cost was an observation that ignoring out-of-bounds ports would be less costly than dealing with them. > (ii). All of the IETF stuff > > softwire, behave and coexist work > http://www.ietf.org/internet-drafts/draft-nishitani-cgn-00.txt Didn't find an instance of the word "cost" in here. Wouldn't expect to in an I-D. But then that's my point: most likely nobody has attempted to do a real cost analysis on deploying and maintaining an Internet where the residential accounts are mostly behind CGNs. A claim that such a system will cause an alarming increase in cost is not well supported. Given the crushing blow to P2P consumption and increased revenue from the sales of static IP addresses, the delta cost could be zero or even negative. I doubt it will be, but I also doubt it will be alarmingly high. Regards, Bill Herrin -- William D. Herrin ................ [EMAIL PROTECTED] [EMAIL PROTECTED] 3005 Crane Dr. ...................... Web: <http://bill.herrin.us/> Falls Church, VA 22042-3004 _______________________________________________ rrg mailing list [email protected] https://www.irtf.org/mailman/listinfo/rrg
