> From: Alia Atlas <[email protected]>
> I want to see something that has the assumptions written down and
> can be implemented.
How about 'Anything that can be fed _into_ the mapping system, and get an
answer, is an LEID; anything that can be seen in the _output_ from the
mapping system is an RLOC.' Although even that's probably too precise - it
my mind were fully awake, I could probably come up with some countexamples.
> At least one case of parts that don't work when the same bit-pattern
> is used as an EID for one host and as an RLOC (or globally routeable
> address) for another is:
> ANYWHERE that a lookup in the mapping database is done to determine
> if an address is an EID or RLOC
Mu. The assumption in that question (that something is _either_ an RLOC
_or_ an LEID) does not hold 100.00000% - although I wish it did! (Leaving
aside the cases of legacy addresses, which are all both at once, of
course.) If you can look up a bit-pattern in the mapping database, and get
back an answer, that bit-pattern has LEID-nature. But that doesn't mean it
can't also have RLOC-nature.
Yes, having that kind of mixed-up semantics a lot is not good; and for the
simple (and hopefully by far and away the most common) model, it does not
happen. But the Internet is large, and there are more kludges in it than any
of us dream of in our architectures....
Noel
_______________________________________________
lisp mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/lisp
