Hi Karsten,
Please see my reply inline [FP], thanks. From: Karsten Thomann [mailto:[email protected]] Sent: Saturday, July 05, 2014 4:23 AM To: [email protected] Cc: Fan, Peng Subject: Re: [sunset4] Fwd: New Version Notification for draft-fan-sunset4-router-id-00.txt Hi Peng, some additional comment to the draft as I now had some more time... The use of the term system id and 32 bit in one sentence is in my opinion a bit problematic as system id is an isis term and the sys id of isis is not 32 bit, but up to 48 bit. [FP] I'll remove "system id" if it seems problematic. I guess the ID we are talking about is most frequently referred to as "router id", so calling it "router id" is enough, although the device running the protocol is not necessarily a router. Also with some care it should be a possible solution to use the last 32bit (or 48 for isis) of the IPv6 address and convert them to decimal to generate the initial ID. I know thats the thing with hex converted "IPs" we already discussed, but this time instead of manuelly created ID as a way to create it automatically. [FP] Yes. We are actually focusing on two aspects: the generation of unique IDs and the possibility of binding ID with IP address. If IP address is planned beforehand, and router ID is not planned but left to admins to decide, then router ID can be determined using IP address, e.g. using hash as described in the document, or directly copying part of IP address as you have suggested (perhaps we can consider it as a simple hash). But it is possible that the same part of two IP addresses are the same, so we have to take care when doing this. Thanks and regards, Peng Regards Karsten Am Freitag, 27. Juni 2014, 23:00:44 schrieb Fan, Peng: > Hi Karsten, > > Okay so it is inconvenient at least when we want to get the value of router > id from an address. > > This discussion does remind me of a saying that mixed notation is not > suggested... http://www.ietf.org/mail-archive/web/idr/current/msg12901.html > > Best regards, > Peng > > > -----Original Message----- > > From: sunset4 [mailto:[email protected]] On Behalf Of Karsten > > Thomann > > Sent: Friday, June 27, 2014 5:24 PM > > To: Fan, Peng > > Cc: [email protected] > > Subject: Re: [sunset4] Fwd: New Version Notification for > > draft-fan-sunset4- > > > router-id-00.txt > > > > Hi, > > > > The mixed notation can't be used, as it is converted to the hex version > > automatically and is stored as the /96 case in hex. > > ping 2001:db8::192.168.0.1 > > Pinging 2001:db8::c0a8:1 with 32 bytes of data: > > > > At least im not aware of a device which stores the IPv6 address in mixed > > notation, they accept it and convert it to the hex notation in /96, afaik. > > > > Regards > > Karsten > > > > Am 27.06.2014 10:43, schrieb Fan, Peng: > > > Hi Karsten, > > > > > > Thanks for the comment, please see reply inline [FP]. > > > > > >> I think the solution area should be extended by another possibility > > >> to > > > > > > embed > > > > > >> IPv4 addresses in v6, as the /96 version has to use the hex version > > >> of the > > > > > > ID, > > > [FP] Perhaps we can use the mixed notation, e.g. > > 2001:db8::192.168.255.1? > > > >> the use of a /80 could be used to store them in a format like > > >> 2001:db8::1921:6825:5001 or 2001:db8::0102:5525:0001. > > >> If you like to use an even better readable approach you can also use > > >> a /64 > > > > > > for > > > > > >> addresses like 2001:db8::192:168:255:1. > > > > > > [FP] Yes these approaches are quite human-friendly, thanks. > > > [FP] We can also consider if there is another solution besides address > > > mapping and other ideas in the draft. > > > > > > Best regards, > > > Peng > > > > > >> Regards > > >> Karsten > > > > > > _______________________________________________ > > > sunset4 mailing list > > > [email protected] > > > https://www.ietf.org/mailman/listinfo/sunset4 > > > > _______________________________________________ > > sunset4 mailing list > > [email protected] > > https://www.ietf.org/mailman/listinfo/sunset4 > > _______________________________________________ > sunset4 mailing list > [email protected] > https://www.ietf.org/mailman/listinfo/sunset4
_______________________________________________ sunset4 mailing list [email protected] https://www.ietf.org/mailman/listinfo/sunset4
