Ian,
I've been meaning to write a draft for the longest time that has some
overlap here; maybe you can save me some work ;) When I was working
elsewhere, we asked NetApp to allow some additional control over proxies.
Basically, it was to allow the following operations (IIRC some of this was
our idea, some was already there):
* marking an object stale
* ejecting an object from cache
* prefetching an object into the cache
(the third needs some explanation - basically, when directed, the cache will
fetch the object as normal, but will not return an object body in the
response, just the status and headers as normal; useful for seeding caches
with large objects, and not sending the object round-trip).
The mechanism that made sense at the time was to use Cache-Control request
headers instead of specialised methods, e.g.,
GET http://www.foo.com/bar.gif HTTP/1.0
Cache-Control: prefetch
I still think this might be a better/more natural way to communicate these
directives, rather than using a new HTTP method. Thoughts?
BTW, in the tracecodes, perhaps it would be better to use 'VALIDATED' rather
than 'VERIFIED', to stay in line with HTTP language?
Also, is it a good idea to use 2774, considering that it's experimental?
Cheers,
On Wed, May 31, 2000 at 04:29:17PM +0100, Ian Cooper wrote:
> Apologies for the noise to those that have seen this before; I was
> asked to send the attached to the list after the Web Caching and
> Content Distribution Workshop in Lisbon. I guess the main interest
> was in the sections on purging content
>
> This is an unedited copy of the original submission (almost a year
> old), so the contact information is now out of date. Send me email if
> you need the new details for some reason.
>
>
> One comment that I'll add (which was suggested to me, and is included
> in the draft): there was a lot of talk about "new HTTP headers" at the
> workshop. We should probably all make an attempt to use the HTTP
> extension framework (RFC2774)
-- Mark Nottingham, Senior Developer Akamai Technologies (San Mateo, CA)
This archive was generated by hypermail 2b29 : Thu Nov 18 2004 - 11:21:28 MST
