W3C home > Mailing lists > Public > ietf-http-wg-old@w3.org > September to December 1996

Re: Hit-metering: to Proposed Standard?

From: Ari Luotonen <luotonen@netscape.com>
Date: Wed, 20 Nov 1996 07:14:00 -0800 (PST)
Message-Id: <199611201514.HAA13804@step.mcom.com>
To: Benjamin Franz <snowhare@netimages.com>
Cc: http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com

Benjamin Franz wrote:
> No - I think that after some thought, most people would conclude that they
> *want* that information, but are not necessarily *entitled* to that
> information. The key here is that they are paying their *service provider*
> not *us*.

They are paying for their online service provider for their internet
connection, but they are not paying their online service provider for
the services rendered by origin servers, nor is the online service
provider paying those origin servers for their services.


> The *obligations* are therefore between them and their service
> provider - not between them and the net at large.

If this is the case, this means that there is a corresponding
obligation between the service provider and the origin server
(otherwise you could get total immunity, and no one would be liable to
the origin server), which brings us to the issue at hand, and it *is*
the service provider's responsibility to report back accesses to the
origin server.


> This is an important
> distinction. The relevant model is caller ID to my mind. 
> 
> Businesses pay for phone lines so that they can communicate with their
> customers. Many would *like* to identify their customers phone numbers. 
> But they are not *entitled* to it, and I can block caller ID and something
> like 50% of people in California do have full blocking.

The Caller ID analogy is not relevant to the hit reporting draft.  The
hit reporting draft doesn't give out any information about the
requesting user/client, only the fact that it was requested.

Phone usage billing might be a closer analogy.  The origin servers are
only given enough information so they know how many hits they really
got -- that is, what is the quantity of their services used.


> Californians value their privacy. 

Before there's any confusion on this matter: the hit reporting in
Mogul, Leach draft discloses ONLY the number of hits, NOTHING else.
That's the least bit of information you need to have in order to find
out how much of your services were used.  No individual users' private
information was disclosed.


> Am I missing something here? Why would large online services give *any*
> information about their proxy stats to an outside group? I certainly would
> not do so for Joe Q. Not My Customer.

Because you _are_ relaying _their_ services, for _your_ customers,
your paying customers that have chosen to use that service.  By
co-operating you can serve that data faster from your cache, and
you're not "stealing" the data and making your own illegal copies.


Cheers,
--
Ari Luotonen	* * * Opinions my own, not Netscape's * * *
Netscape Communications Corp.		ari@netscape.com
501 East Middlefield Road		http://home.netscape.com/people/ari/
Mountain View, CA 94043, USA		Netscape Proxy Server Development
Received on Wednesday, 20 November 1996 09:36:19 EST

This archive was generated by hypermail pre-2.1.9 : Wednesday, 24 September 2003 06:32:16 EDT