Keith Moore wrote:
: IPv6 with unnecessarily lengthy 16B addresses without valid
: technical reasoning only to make network operations prohibitively
: painful is a garbage protocol.
:
: LISP, which perform ID to locator mapping, which is best
: performed by DNS, in a lot less scalable way than DNS
: is a garbage protocol.
Perhaps your feedback would be more generally useful if you described,
in technical terms, why you believe DNS is a better ID-to-locator
mapping than LISP is.
You should have asked so in 2019.
Anyway, with my reconfirmation in 2021, there was discussion on IPv6
which, I believe, concluded that using MAC address as part of IP
address is layer violation because MAC address length affect
IP address format.
But, in 2021, seemingly, no one was serious for LISP.
(I personally see a lot of technical problems with using DNS as an
ID-to-locator mapping protocol, due to lack of speed, security, and
reliability if not very robustly provisioned.
As for speed, it's IP mobility which should be responsible for
it. For security, I will post another message for a separate
thread.
> I'm not even convinced
> that an ID/locator separation is desirable.
Destination locator rewriting, for example, is useful for IP
mobility forwarding to avoid IP over IP tunneling, that is, not
to reduce MTU by rewriting home locator to foreign locator.
Anyway, such rewriting is not possible with LISP where transport
checksum (knowledge of which is unavailable at IP layer) is
invalidated by the rewriting.
Masataka Ohta
