Re: [lisp] I-D Action: draft-ietf-lisp-eid-block-mgmnt-01.txt

[Geoff Huston <[e-mail]>]

On 15 Feb 2014, at 3:34 am, [e-mail] wrote:

> 
> A New Internet-Draft is available from the on-line Internet-Drafts directories.
> This draft is a work item of the Locator/ID Separation Protocol Working Group of the IETF.
> 
>       Title           : LISP EID Block Management Guidelines
>       Authors         : Luigi Iannone
>                         Roger Jorgensen
>                         David Conrad
> 	Filename        : draft-ietf-lisp-eid-block-mgmnt-01.txt
> 	Pages           : 13
> 	Date            : 2014-02-14



section4, bullet 5 states:
      All allocations (renewed or not,
      including delegations and sub-allocations) MUST be returned by 31
      December 2020, in accordance to the 3+3 years plan outlined in
      [I-D.ietf-lisp-eid-block].


but the text at the end of the section reads:

  If/When the EID block experiment changes status (e.g., to not being
  "experimental"), and following the policies outlined in [RFC5226],
  the EID block will change status as well and will be converted to a
  permanent allocation. 


Bullet 5 states "MUST be returned" and the later text states "will be converted to a permanent allocation"

This seems to be a contradiction. What's the intended plan? 

If the permanent plan is that LISP runs from corralled space, then I am seriously concerned that this is an admission of failure of LISP from the outset. I though the object of the exercise was to offer LISP as a routing protocol with superior scaling properties to what we have now. But if this entails renumbering the Internet to achieve it, then just renumbering the Internet so that the address structure aligns with the topology of the network would allow the existing protocols to also scale - so where is the "win" in LISP?

At the very least it would be good for the draft to clarify the directives of must be returned and the conversion to a permanent allocation.

But I would also like to understand the longer term issues at play here - is the longer term plan for LISP to route the Internet's unicast address space as deployed, or are we truly contemplating a lengthy transition into an essentially renumbered space?

Geoff