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

Re: Hit-metering: to Proposed Standard?

From: Gregory J. Woodhouse <gjw@wnetc.com>
Date: Mon, 25 Nov 1996 16:30:02 -0800 (PST)
To: Jeffrey Mogul <mogul@pa.dec.com>
Cc: Ingrid Melve <Ingrid.Melve@uninett.no>, http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com, http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com
Message-Id: <Pine.SGI.3.95.961125161852.18540C-100000@shellx.best.com>
In the case of URNs that resolve to multiple URLs, it seems the answer
depends upon the purpose of hit metering. If it is a server management
tool, then the most important information is the number of hits to a given
resource (meaning something identified by a URL). If the point i to measure
the usage level of a URN-resource (Surely, there must be some established
terminology here!) then it seems appropriate to meter the accesses to the
URN. This is problematic, though. Under the NAPTR proposal (which, of
course, is only one possible approach), then the question arises of how to
handle N2L or N2Ls responses, where each URN-request results in a URL
request (terminology?), but with N2R responses, the actual resource could
potentially be served by any of a number of protocols.

It seems to me that hit metering (restricted to HTTP) and resource usage
(where a resourcer is more broadly defined as something identified by a
URI which needn't necessarily be an http: URL) are different problems. 

---
Gregory Woodhouse     gjw@wnetc.com
home page:            http://www.wnetc.com/home.html
resource page:        http://www.wnetc.com/resource/
Received on Monday, 25 November 1996 16:42:08 EST

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