Re: Accounting in peered CDNs: (User Profiles)

[Oliver Spatscheck <spatsch@research.att.com>]

Actually the user profile I was looking for is more an aggregate of
who watches the site (slow reporting time scale of minutes, hours, days..)
What you are talking about is the use of real time profiles which seems
also valid, but different.

Oliver


Michael Vernick writes:
 > User profiles are very important, especially if you include profile information
 > about the user's device to make decisions on the delivered format of content.
 > i.e. content created for a standard PC screen could be modified within the
 > network for a device with a small screen.
 > 
 > Have you looked into the work going on in the CC/PP working group in the W3C?
 > (http://www.w3.org/Mobile/CCPP/)  This may be what part of what you are looking
 > for.
 > 
 > Michael Vernick
 > Lucent Bell Labs
 > 
 > Reinaldo Penno wrote:
 > 
 > > Hi,
 > >
 > > I was thiking about this problem (stats for demographics) some time ago and
 > > have been working on a Used Profile Protocol. The abstract goes like
 > > this...(I'm not sending the whole text because do not want bother the
 > > members on this list if this is not the right forum)
 > >
 > > ---
 > >
 > > Network Working Group                                           R. Penno
 > > Internet-Draft
 > > NortelNetworks
 > > Expires: June, 2001                                         December, 2000
 > >
 > >                       User Profile Information Protocol
 > >                     draft-penno-cdnp-nacct-userid-00.txt
 > >
 > >    We present here a protocol in which edge network devices, which can be
 > >    further be subdivided in  IP Services Boxes, Broadband RASes, Edge
 > > Routers
 > >    and so forth, can inform surrogates or traffic interception devices
 > > extended information
 > >    about the user, such as geographic location, QoS policy, login (
 > >    name@domain.name), and start and stop times for the session.
 > >
 > >    This would not only allow better statistics and billing but also provide
 > > a way to
 > >    to surrogate servers and traffic interception devices to treat
 > > subscribers differently,
 > >    based on this extended information
 > >
 > > ---
 > >
 > > would this working group be the right forum for this protocol?
 > >
 > > thanks,
 > >
 > > you# -----Original Message-----
 > > you# From: owner-cdn@ops.ietf.org
 > > you# [mailto:owner-cdn@ops.ietf.org]On Behalf Of
 > > you# Oliver Spatscheck
 > > you# Sent: Wednesday, January 17, 2001 4:00 PM
 > > you# To: Abhi Deshmukh
 > > you# Cc: cdn@ops.ietf.org
 > > you# Subject: Accounting in peered CDNs: Initial thoughts ....
 > > you#
 > > you#
 > > you#
 > > you# Abhi,
 > > you#
 > > you#  great summary. One thing we always hear from content
 > > you# provider is the
 > > you# request of stats for demographics (at least location if possible
 > > you# more). If we therefore, envision a world in which the
 > > you# entire content of
 > > you# a site would be delivered using CDN peering being able to aggregate
 > > you# client demographics would be one requirement. However, I recognize
 > > you# the complexity of that problem and I am not sure if we
 > > you# want to address
 > > you# it in the first version.
 > > you#
 > > you# Oliver
 > > you#
 > 
 >