Re: [lisp] WG Last Call on draft-ietf-lisp-eid-block-mgmnt-02

[Geoff Huston <[e-mail]>]

As promised, here are some comments on Section 5

1. replace "Allocation" with "Registration"

As I noted in the last note, the terms "allocation" and "Assignment"
are hopelessly overloaded, and best avoided, as the distinction
is really irrelevant here,


    5.  EID Prefix Registration Requirements

2. are you talking about the registry or the registered end users?

reword the section as folllows:


   All EID prefix registrations should be made according to the
   following technical constraints:


   1.  All EID prefix registrations MUST use a globally unique EID prefix.

   2.  If there is more than one registry operator, all operators MUST use the
       same registry management policies and practices.

   3.  The EID Prefix registration information as specified in Section 6, MUST be collected upon initial registration and renewal, and made publicly
       available though interfaces allowing both retrieval of specific registration details (search)
       and enumeration of the entire registry contents.

(enumeration is a tricky topic - in an experimental context I think it helps with the experiment, but
your mileage may differ)

I suggest dropping 4 - it heads into over-specification, and this is not a services contract nor a SLA specification.

|   4.  The registration information service should be reasonably
|       reliable so to make such information readily available.  The
|       allocation service SHOULD be provided during regular business
|       hours in venue in which the allocation service is housed.


5 is ambiguous - do you want reverse or not? Lets say "yes"


   4.  The registry operator MUST permit the delegation of EID prefixes
       in the reverse DNS space to holders of registered EID prefixes.

6 is the "non-discrimination" clause isn't. Anyone can obtain a registration, on the understanding that the
prefix so registered is for the exclusive use in the LISP experimental network, and that their registration
details (as specified in section 6) are openly published in the EID prefix registry.

So why not say that?

    5. Anyone can obtain an entry in the EID prefix registry, on the understanding that the
       prefix so registered is for the exclusive use in the LISP experimental network, and 
       that their registration details (as specified in section 6) are openly published in the
       EID prefix registry.



Geoff