Re: [Casm] I-D Action: draft-kumar-casm-requirements-and-framework-00.txt

Hi Brian,

Thanks a lot for review and comments.
I apologize for missing your previous message (not sure if I saw it). Anyway, we had some private discussions before this email alias was created, some folks wanted to keep the word “centralized” to emphasize that a central entity should be responsible for managing network addresses. One of the benefit is single pane of glass that can be used for other use cases such as tracking a device/user through the network in real-time. 

But after seeing your comments, I am beginning to think that may be not all aspects of address management need to be centralized. If we were to develop an address management service, what can/should/must be centralized and what should be left up to implementations. I think it should be a good topic for debates and discussions and then decide how to address this in the draft.

I would really appreciate your help in driving this draft.

Thanks 
Rakesh
On 2/7/17, 7:40 PM, "CASM on behalf of Brian E Carpenter" <casm-bounces@ietf.org on behalf of brian.e.carpenter@gmail.com> wrote:

    Hi,

    > This document describes evolution of IPAM into a
    > standardized interfaces for centralized management of IP addresses.

    > CASM:   Centralized Address Space Management

    Please delete 'centralized'. It isn't necessary (as mentioned in my
    previous message) and it has the effect of limiting the range of
    solutions that can be considered. I will consider the document as
    if this word does not appear...

    And in fact I like the draft. I simply observe that this part of the model:

            +---+------------+-------+-------+---------------+------+
            |    Address Space Management (IPAM) System             |
            |      +-----------+ +----------+ +--------+            |
            |      | Pool      | |Address   | |Database|            |
            |      | Management| |Management| |        |            |
            |      +-----------+ +----------+ +--------+            |
            |                                                       |
            +-------------------------+-----------------------------+

    *does not need to be centralized* except for two points: providing
    the initial pool to the distributed system, and providing policy
    intent to the distributed system.

    And there's a draft describing a distributed solution, for a subset
    of the problem space:
    https://tools.ietf.org/html/draft-ietf-anima-prefix-management-02
    I understand that it's too soon for this list to discuss solutions,
    but I really do want the distributed approach to be on the table, since
    this is one of the early use cases for the Anima WG.

    Regards
       Brian Carpenter

    _______________________________________________
    CASM mailing list
    CASM@ietf.org
    https://www.ietf.org/mailman/listinfo/casm