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

Sarr Blumson (
Mon, 22 Aug 1994 15:19:45 -0400

Rob Raisch, The Internet Company, says
On Mon, 22 Aug 1994, Sarr Blumson wrote:

> Let's see how this goes if we substitute "book store" for "caching server"

Ummm, please excuse my flipness, but your argument is specious.

A bookstore cannot provide my content -- for no cost -- to thousands
(perhaps millions) of consumers.

No. A book store can only provide your content at to thousands of consumers
at greatly reduced cost, by not paying you.

> My flipness aside, there are issues here, but they are issues which are
> amenable to technical solutions. There are lots of authentication mechanis
> for insuring that a publisher only "sells" to a "distributor" (cache) that
> trusts, and to sign time critical material to that it identifiable as relia
> as a newspaper is by the date on the top. If anything, our problem is
> choosing among the alternatives, which is a problem largely because the
> choosing may decide a lot of things about who get rich from this new
> technology.

I'm sorry, but I strongly disagree with you. The problems are not
amenable to technical solutions -- in the absence of the publisher.

The real problem with the Web and Mosaic and all the rest is that these
are publishing "solutions" designed by technologists.

Go and tell a publisher that everything they have known about control and
intellectual property is now wrong.

I'm not talking about doing anything in the absence of publishers, nor am I
talking about telling publishers that what they are doing is wrong. On the
contrary, what I'm saying is that electronic distribution requires middle men
just as much as paper distribution does, and that we're creating problems
where none exist because we call the middle men distributors for paper and
caches for bits.

John Labovitz <> suggested that what he needs is an accounting of
what the cache/distributor has done. Providing that is technically quite
feasible. If you're charging for the materials than sending a bill is too, as
is refusing to send materials to anyone you can't identify. Yes, you can't
guarantee that someone pretending to be an individual won't won't
redisstribute your material, but a paper publisher can't guarantee that
someone ordering a single copy of a book isn't a copy shop either.

In case it's not clear, I'm not claiming that all of this is in place with
existing protocol definitions, only that they are problems we know, in
principle, how to solve. My concerns are that we not fall into the trap of
not doing anything because our solutions aren't any better than what's already
there, and because an electronic system without middle men won't work any
better than a system where every book reader has to deal directly with the
publisher would.

Sarr Blumson
voice: +1 313 764 0253 FAX: +1 313 763 4434
CITI, University of Michigan, 519 W William, Ann Arbor, MI 48103-4943