Re: [RAM] Re: draft-bagnulo-lisp-threat-01

[Robin Whittle <rw at firstpr.com.au>]

Hi Dino,

You wrote of Marcelo's comment:

>> i don't know, there is no ID describing LISP v2,3,4 afaik
> 
> There will probably never be a LISP 2 because many people have sounded
> off about the circular dependency even though DNS gives the best low
> latency for a request based mapping lookup (that is, it shares the DNS
> lookup latency with the map lookup).
> 
> The LISP authors have never mentioned or defined a variant numbered 4.
> 
> And for LISP 3, there is currently 3 examples of LISP 3, that is CONS,
> NERD, and APT.

I mentioned LISP 2.x, 3.x and 4.x, which are listed in your
message on 16 April:

  http://www1.ietf.org/mail-archive/web/ram/current/msg01289.html

including several 4.x variants.  I don't recall any other mentions
of 4.x.

I don't think there is a problem with circular dependencies with
LISP using DNS provided the LISP-mapped address space comes with a
caution or something to say it shouldn't be used for some DNS
functions, including root nameservers and a few other things, and
any nameserver on which LISP depends.  Brian Carpenter made the
original circularity critique, I think, but on 28 June:

  http://www1.ietf.org/mail-archive/web/ram/current/msg01588.html

I think he agreed to a certain extent with the limitations I
suggested which would prevent circularity problems.  His remaining
critique is that the list of constraints is "operationally error
prone", but I think a lot of things we have to rely on are
"operationally error-prone".


  - Robin


_______________________________________________
RAM mailing list
RAM at iab.org
https://www1.ietf.org/mailman/listinfo/ram