> -----Original Message----- > From: Rémi Després [mailto:remi.despres@xxxxxxx] > Sent: Thursday, September 29, 2011 3:13 AM > To: Dan Wing > Cc: 'Softwires-wg'; 'Behave WG'; 'IETF discussion list'; 'Teemu > Savolainen' > Subject: Re: [BEHAVE] ... Dual Stack Hosts Using "Bump-in-the-Host" > (BIH)) > > > Le 29 sept. 2011 à 23:50, Dan Wing a écrit : > > >> -----Original Message----- > >> From: Rémi Després [mailto:remi.despres@xxxxxxx] > >> Sent: Thursday, September 29, 2011 5:14 AM > >> To: Softwires-wg > >> Cc: Dan Wing; Teemu Savolainen; Satoru Matsushima; IETF discussion > >> list; Behave WG > >> Subject: Re: [BEHAVE] ... Dual Stack Hosts Using "Bump-in-the-Host" > >> (BIH)) > >> > >> Hi all, > >> > >> Le 27 sept. 2011 à 21:10, Dan Wing a écrit : > >> > >>>> -----Original Message----- > >>>> From: teemu.savolainen@xxxxxxxxx > [mailto:teemu.savolainen@xxxxxxxxx] > >> ... > >>>> I mean does existing > >>>> applications work better if double translation is done in > >> deterministic > >>>> manner? > >>> > >>> Yes, it allows the CPE to implement an ALG -- if an application > needs > >>> an ALG (e.g., active-mode FTP). > >> > >> +1 > >> > >> As Softwire is concerned, it is worth noting here that, with > >> encapsulation rather than double translation, NO ALG is ever needed. > >> (Neither in ISP Border Relay nodes, nor in CPEs, nor in BIH hosts). > > > > Do a message flow for active mode FTP with 4rd. > > Two cases (which I din't differentiate, sorry for that). > a) The hosts handles 4rd > b) The host is behind a NAT44 ( > > In case b), an ALG is needed as usual, as the CPE-NAT44 does its usual > job for that. > Case a) was assumed to apply to this discussion, considering a host > that is directly the 4rd CE. > > > I did one, and > > it needs an ALG in the 4rd NAPT44 function. > > In case a), the host can assign to an FTP application a pair of ports > of its assigned range. > Thus no translation is needed (e2e transparency). > > OK? If BIH worked that way, BIH would fail when the host has multiple interfaces running, and those interfaces have different port ranges. -d > RD > > > > > > > -d > > > >> It is sometimes argued that double translation could be as simple > than > >> encapsulation. > >> AFAIK, this discussion clearly indicates the contrary. > >> (No ALGs eliminates any variants about where to put them.) > >> > >> RD > > > > > > _______________________________________________ > > Ietf mailing list > > Ietf@xxxxxxxx > > https://www.ietf.org/mailman/listinfo/ietf > > _______________________________________________ Ietf mailing list Ietf@xxxxxxxx https://www.ietf.org/mailman/listinfo/ietf