Re: CDI Distribution Draft

[Dan Li <lidan@cisco.com>]

Hi, Lisa, I think section 5.5 about relationship with RUP is largely inline 
with what I expected. Some comments about the ones flagged as still lacking 
in RUP:

>In addition to the metadata fields which were defined for CI-DS, but
>    not mapped to an RUP field, CI-DS requires the following
>    functionality.
>
>       Ability to specifically request a Content Set be added or
>       withdrawn from the list of objects being serviced. This
>       requirement could be covered by adding the ability to dynamically
>       define Resource Groups to RUP.

This is a no-brainer. Any grouping protocol got to have it. If you find it 
missing in the RUP draft, it's an editing error. I will add it.

>       Specifically, a master Resource Group could be defined which
>       would be used as the channel to communicate which Resource Groups
>       to be added/withdrawn from service.  The CIG would then establish
>       channels to receive cache coherency messages for these Resource
>       Groups.

Yah, this is a nice architecture point for an end-to-end CDN or inter-CDN 
system. Good to mention it in your draft.

However, it is out side of RUP scope. RUP defines the operation within one 
channel. How multiple channels are tied together or communicated to a cache 
is a separate matter, and can potentially use any one of many solutions.

>       Ability to update metadata for an object.  While it is possible
>       this can be covered by withdrawal and re-add of the Content Set,
>       due to the number and types of fields supported, inability to
>       update would impact efficiency.

Could you explain this a bit more? What kinds of "metadata"? HTTP metadata 
or CDN metadata or something else?

If it's just arbitrary stuff one can associate with a Resource Group, RUP 
allows it. RUP has the "option" fields for people to stick anything there 
as long as the receiving party is equipped to interpret the options.

But if it's some specific metadata and you expect interoperability among 
independent implementations (and operators), then please specify them and 
RUP can look into incorporating them.

Thanks!
Dan

At 06:10 PM 11/19/2001 -0500, Lisa Amini wrote:
>Hi Dan,
>Sorry I've not been able to spend any time on the RUP document, I've been
>completely buried lately.  Anyway, I tried to update the CI Distribution
>Req'ments draft to reflect the work being done in RUP.  I would really
>appreciate if you could review to ensure I have properly understood the
>scope and intent of RUP and WCIP.
>
>Thanks,
>Lisa
>---------------------- Forwarded by Lisa Amini/Watson/IBM on 11/19/2001
>06:08 PM ---------------------------
>
>Lisa Amini
>11/19/2001 06:07 PM
>
>To:   cdn@ops.ietf.org
>cc:
>From: Lisa Amini/Watson/IBM@ibmus
>Subject:  CDI Distribution Draft
>
>
>I've attached the latest version for the CI Distribution Requirements draft
>-- updated based on input from/since the last IETF meeting.  I would
>appreciate any comments/suggestions you may have.
>
>Thanks,
>Lisa
>  (See attached file: draft-amini-cdi-distribution-reqs-02.txt)
>