Re: HTML Objects.

James R Grinter wrote:
>
> On Mon 2 Dec, 1996, Alessio Bragadini <alessio@unipi.it> wrote:
> >HTTP/1.1 does, by Content-MD5 header: it's a MD5 digest of the
> >object. That leads to wonder when (and if) Squid should go to be fully
> >1.1-aware.
>
> but how is the client expected to get that for an unseen before URL,
> without getting the whole thing? HEAD? We all know why IMS queries
> were so popular...
>
> Some form of conditional MD5 GET would make it easier to know if an
> object has changed, maybe, but that's really about it.
>
> URNs really are essential to solve this problem of duplicated objects,
> I think.

If I follow HTTP/1.1 correctly (disclaimer: I may be misinformed and
talking out of my nether regions inadvertantly) the persistant
connection allows us to send several requests and responses without the
overheads of starting a new connection. Therefore it would be reasonable
to do a HEAD to get the MD5, do a lookup in a database to see if we
already have something matching that description, and if not, do a GET
for the object. The RTT on the system would be pretty good, all things
considered, since we don't have to deal with TCP slow-start more than
once.

Once we found an MD5 that matched an existing one, but for a different
URL, we would have to link the URL to the MD5 so that we know we have it
in the cache. Obviously this is something of a jump, but we may be able
to hook it in in the front-end of the code that utilizes the redirector
for simplicity.

D
Received on Mon Dec 02 1996 - 18:47:54 MST