Re: Caching Servers Considered Harmful (was: Re: Finger URL)

James 'J' Allard (
Tue, 23 Aug 1994 02:55:20 +0200

| > One solution would be for caching servers to generate a summary of
| > hits on URLs `belonging' to particular servers, and to email that
| > summary to a standard email address at those servers.
| Better might be to add a method that passes on statistics. Eg,
| CACHERECORD /ads/citycorp/index.html 52
| meaning that 52 requests for /ads/citycorp/index.html were served from
| the cache.
| The biggest problem for me as a server administrator is that caches
| hide accesses and popularity, and that's something that I *need* to
| have, to justify the cost of the server. I'd like to see that change.

although polite, unlikely that this issue will gain universal acceptance.
we try to keep stats on, but know that there are at
least 12 unofficial mirrors that deny this, or refuse to give us these
stats by matter of convention. technically, you can't enforce something
like this into the protocol, since it can be easily circumvented and
undetected. i assert it will be if such a mechanism is offered.

that's not to say that i disagree with the proposal, but that your "need"
will only be satisfied by disallowing the caching of documents and
forcing authentication for access. that's the closest you'll get to a
real count, but you'll never be able to find those that pirate your info.
sure, a caching server can ignore your request not-to-cache, but it
will hose the stock ticker service in the process, and likely will be
controlled by the users eventually.

J. Allard
Program Manager of TCP/IP Technologies work: (206)882-8080
Microsoft Corporation home: (206)860-8862
"On the Internet, nobody knows you're running Windows NT"