I just remembered. It was so the LSBs map consistently to the RLOCs in the RLOC-set across all ETRs. So for example, if the RLOC-set was like this in each Map-Reply:
>From ETR-A: RLOC = (1.1.1.1, 2.2.2.2, 3.3.3.3) (i.e. all are priority 1) >From ETR-B: RLOC = (3.3.3.3, 2.2.2.2, 1.1.1.1) (i.e. all are priority 1) Then when an ITR cached the ETR-A order when ETR-A had an LSB of 0x6 (b'110) it is conveying that 1.1.1.1 is down. If ETR-B returned a Map-Reply and also believed that 1.1.1.1. was down, its LSB setting would be 0x3 (b'011). Each, in its own right are encoding the LSBs consistent and it does not really confuse the ITR as long as the ITR does not encap to 1.1.1.1. The ITR, also acting as an ETR would have to know then A and B are encaping that the LSB is changing values but not changing in semantics. I was assuming here that ETR-A is 2.2.2.2 and ETR-B is 3.3.3.3. Dino On Oct 3, 2014, at 7:58 AM, Dino Farinacci <[email protected]> wrote: > That was going to be my first deep-thinking about how to add text. And if it > was really necessary. Back when we first designed the RLOC-set, we wanted > them sorted to be sinked between ETRs that each would be sending Map-Replies. > But what matters is the order of priorities. It allows one to reduce scans > for the best priority RLOCs if they are assumed to appear at the beginning of > the list. > > I think there were other reasons. I'll go through my design notes from 2007. > > Dino > > On Oct 3, 2014, at 7:35 AM, Sander Steffann <[email protected]> wrote: > >> Hi, >> >>> We have seen that, according to RFC 6830, the locators appearing in a Map >>> Reply must be "sorted in order of ascending IP address where an IPv4 >>> locator address is considered numerically 'less than' an IPv6 locator >>> address". >> >> Seeing this whole discussion I started to wonder: why must locators be >> sorted anyway? >> >> Cheers, >> Sander >> >> _______________________________________________ >> lisp mailing list >> [email protected] >> https://www.ietf.org/mailman/listinfo/lisp > _______________________________________________ lisp mailing list [email protected] https://www.ietf.org/mailman/listinfo/lisp
