Le 8 nov. 2011 à 11:35, Mark Townsley a écrit : > > On Nov 8, 2011, at 10:12 AM, Rémi Després wrote: >> >> - Talking with Mark townsley, I got the understanding that this wasn't a >> real problem, at least with IOS. >> => Clarifying this point would IMHO be useful. > > Our tunnel code typically processes tunnel packets upon receipt at a single > address (/128 for v6 or /32 for v4). This is by far the easier and more > "normal" way of doing things for the myriad of tunnels we support today. > Implementing on receipt of anything other than an address (or hostroute, if > your prefer since this is a router) is "possible" but would certainly be > considered "special" and as such has a higher bar for implementation within > the typical tunneling architecture (at least in the implementations I am > aware of, YMMV). > > Apologies if I was not clear on this point when we talked the other day.
Apologies if I took what you said more positively (IMHO) than what you wanted to express. I am still curious to see how NAT64 is implemented, including in open-source code. However, I am fine with leaving aside for the time being this question of code complexity (as you say subject to YMMV), and with focusing on what operational features depend on this capability. Cheers, RD > > - Mark
_______________________________________________ Softwires mailing list [email protected] https://www.ietf.org/mailman/listinfo/softwires
