- From: Jeffrey Mogul <mogul@pa.dec.com>
- Date: Tue, 26 Nov 96 10:54:51 PST
- To: hardie@nasa.gov
- Cc: http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com
Sorry for the slow response ... > If someone is able to describe a specific scenario where the use > of the Meter mechanism, as proposed in our draft, does in fact provide > more per-client information than the existing HTTP/1.1 mechanisms, > then we would regard this as a bug in our specification that needs > to be fixed (or at least, that needs to be called out in the Security > Considerations section). The use of the Vary header in a do-report situation clearly provides more information than is currently the case where a proxy cache is being used. Currently, if I employ a proxy-cache and it requests a resource on my behalf, the origin server gets the data on the proxy cache (the cache may report through some data on the origin requestor, but it doesn't have to). If the origin server cache-busts, the proxy-cache must re-request the data every time, but the origin server gets the data on the proxy-cache every time. With your proposal, it could get aggregates of the data on the actual requestors. This compromises privacy. If you are comparing the "Meter: do-report" situation with the one for a fully cachable response, then, yes, the origin server does get more information. However, if you would instead compare the "Meter: do-report" situation with the "Cache-control: max-age, must-revalidate" situation THAT IS ALREADY ENABLED BY THE HTTP/1.1 SPECIFICATION (sorry for shouting), then I do not believe it is possible for the origin server to obtain more data using "Meter: do-report" than it could without it. In fact, because the hit-metering mechanism *does* aggregate data regarding multiple requests (and probably multiple clients), it actually delivers *less* data to the origin server than would be the case if the origin server did simple cache-busting. I.e., the origin server would see the count of the number of clients who preferred to see their documents in Kurdish, but not the actual request headers. I view this as a potential improvement for privacy (although any actual improvement clearly depends on the goodwill of the origin server). Note that this *reduction* in the data is precisely what Phill Hallam-Baker does not like about our proposal. You wrote: Imagine for a moment that someone used a Vary: on the Host header with Meter. and then followed that up with Of course the host header would tell you nothing about the user. Imagine some other header instead. We did try to "imagine some other header", and could not find any specific example of a request header (or any other piece of information) that the Meter mechanism would allow the origin server to obtain that was not otherwise obtainable using the features of HTTP/1.1. -Jeff
Received on Tuesday, 26 November 1996 11:17:37 UTC