Posted by David Dorward on 06/08/05 00:26

Luigi Donatello Asero wrote:

> As you snipped the google URL I cannot answer your question.

Why not? I left the references header intact and you posted the message in
the first place.

> But I suppose that it had to do with that https pages can be cached.

Why did you feel you needed to make that point?

Caching is an interesting subject. There are basically two reasons why
something may be cached.

1. By a client, so the next time the same user visits the page the data can
be retrieved locally and thus save time and bandwidth.

2. By a proxy in the middle, so that next time a different user of the same
proxy requests the page, it can be retrieved locally (for the same
reasons).

If the page is transmitted over SSL, then it is encrypted between client and
server. As far as I know, any proxy won't get to see the data in clear
text, so it can't cache it for another local user. (I don't know for sure,
and I suppose its possible that the client requests the page from the proxy
using the proxy's encryption key, and then it requests the page from the
server, decrypts it, reencrypes it then sends it back to the user - however
this would require that the proxy be truely trusted, so I doubt that it is
what happens).

Google is an somewhat special case. It caches pages for (I assume) reason 1.
As an additional service it makes its local cache public. This it can do
because it never sends personal data to the server (since GoogleBot doesn't
have any personal data).

--
David Dorward <http://blog.dorward.me.uk/> <http://dorward.me.uk/>
Home is where the ~/.bashrc is

[Back to original message]