RE: Removal of accounting from scenarios

[Coppola Crescenzo <Crescenzo.Coppola@TILAB.COM>]

Hi team. 

-----Original Message-----
From: Phil Rzewski [mailto:philr@inktomi.com]
Sent: mercoledì 7 novembre 2001 20.31
To: cdn@ops.ietf.org
Subject: Removal of accounting from scenarios


As mentioned in the previous note, in the newly published Scenarios
update
(<http://www.ietf.org/internet-drafts/draft-day-cdnp-scenarios-04.txt>ht
tp:
//www.ietf.org/internet-drafts/draft-day-cdnp-scenarios-04.<http://www.i
etf
.org/internet-drafts/draft-day-cdnp-scenarios-04.txt>txt), I deleted the
previous Section 4, which had details on accounting.

As a first exercise, I'd like to ask that people do an open-minded
reading of
the new Scenarios draft, pretending that you never saw the old Section 4
before. When you're done, are you left with a hollow feeling wondering
about
business models? :)

Ok, that exercise aside, I could put the text back fully intact if
that's
what's desired. However, in trying to make updates to it, I realized the
following things:

1) In the past year, we've almost never made reference to those
accounting
scenarios in our discussions on the list or at the BOFs.

2) The other drafts also make little-to-no reference to the topics
described in
the accounting scenarios section.

3) Some would argue that, because they describe business models, they
don't
really belong in an IETF discussion/draft anyway.

Re: point "3", I do think that some understanding of business models can
help
make the motivation of our work clearer. That being said, I'd like to
propose
that we rebuild any statements we make about business models from the
ground
up, especially in light of the different state of the CDN market and the
economy since a year ago when Scenarios was first drafted.

Recall that the previous draft made reference to the key assumptions:
"Content
Has Value", "Distribution Has Value", and "Clients Have Value". Then, in
what
could be seen as a zoom-in on the "Content Has Value" case, there were
descriptions of payment scenarios: "Cable Scenario", "Telco Scenario",
"Ticket
Scenario", and "Calling Card Scenario".

To provide a straw man of how to replace this as simply/briefly as
possible,
I'd like to instead propose that we fundamentally stress the
"Distribution Has
Value" case. This is to say, the functions provided by a CN (which
really
consists of all of RR/D/A) are, by nature, a service (much like
bandwidth in
the lower-stack universe), and so there's going to be a cost associated
with
that service, regardless of the content flowing across it. This implies
that
there's almost certainly going to be some money flow toward the party
doing the
distribution. I think this is reflected in things like our accounting
drafts,
where our accounting is centered primarily around records that indicate
whether
the enlisted CN did its job: What did you distribute? What requests do
you
route? It's material for the originating CN to look at and determine
"Did I get
the service I paid for?"

I'd like to propose that we simply make the assumption that "Content Has
Value"
will be handled in one place: DRM. This is to say, if the publisher or
originating CN wants to attach rules of use to the content (e.g. style
of
choice: cable, telco, ticket, calling card) they're going to wrap that
into
some DRM mechanism, and they're going to take the responsibility for
collecting
on that. Perhaps the DRM would allow them to delegate authority to
collect to
some other party, who may in fact be another CN. However, what I want to
propose is that such delegation not be a component of CDI (i.e. the
"tree" for
delegation of authority to collect on valuable content should not map
directly
onto the "tree" for delegation of distribution of content).

A third thing we can briefly describe is the ad-revenue model. I think
we all
grok that one pretty well.

With these as a basis, the business scenarios are pretty simple:

Basis: Distribution typically will always have some cost associated with
it,
but anyone enlisting distribution services is going to then have another
strategy for recouping their costs.

1) One way to recoup these costs is to wrap one's valuable content in
DRM to
define rules of use. Therefore, costs are recouped by collecting on fees
for
use of content. The attachment of DRM to content and delegation of
authority
for collecting fees is out of the scope of CDI.

2) Another way is to attach advertisements to one's content. Therefore,
costs
are recouped by collecting fees from third parties that supplied the
advertisements. It's assumed that the accounting data (e.g. usage logs)
collected from an internetworked set of CNs could be used as the basis
for
billing a third-party advertising entity, though how this is done is out
of the
scope of CDI.

What do people think?

--
Phil Rzewski - Senior Architect - Inktomi Corporation                  
650-653-2487 (office) - 650-303-3790 (cell) - 650-653-1848 (fax)