AOL Adds a Cache
Nick Arnett
arnett at alink.net
Thu Apr 16 14:22:08 EDT 1998
At 10:49 AM 4/16/98 -0700, Michael Sauers wrote:
>What about copyright issues in storing these sites en masse without asking
>the permission of the author?
AOL is far from the first organization on the net to use a caching proxy
server. There are thousands and thousands of them, though not too many on
the scale that AOL operates. I seriously doubt if anyone could make a case
for copyright infringement. Caching is a legitimate purpose,
well-established. It is time-shifting, similar to videotaping a television
show so that you can watch it later, which was established as legal in the
Sony decision.
See http://www.inktomi.com/Products/TraffServ.html for information about
the server they're using. It is not a library in any sense. Users won't
even know it's there most of the time.
>What are the guarantees that the user will be
>getting the most up-to-date version of the page if it comes out of the
>cache?
Proxy servers use an HTTP HEAD request to see if a document is up to date.
If the document has changed, it gets the new one and passes it to the user.
There's a lot of intelligence in them, particularly this one. They can
help make the Internet far more efficient by reducing the amount of bits in
at least two ways. HEAD requests are far shorter than full retrievals; if
the server at the other end supports HTTP 1.1, a caching proxy like Inktomi
can check all of the HEADs in one transaction, which is a huge reduction in
packet traffic (there can be a lot of overhead opening and closing TCP
connections in HTTP 1.0). I took a look at these issues a while back
because Inktomi was trying to recruit me to be the product manager for that
server, but this kind of thing doesn't interest me -- too much of a
technically driven product.
Nick Arnett
--
Phone/fax: (408) 733-7613 E-mail: narnett at mccmedia.com
More information about the Web4lib
mailing list