OK so we can only have one cached copy per URL, is that correct? If so the only solution I can think of is by at least using that copy for the main content and once an user requests the Italian version of the content we can always provide the user with non-cached content and non-caching the content, correct? To do so what we have to do is to just add: cache deny itlanguage right? What in any case surprises me is that in the example I gave you requests affected by acl specialuser is always fetching the content from the cache_peer named specialserver despite there is almost always a cached copy of the exact URL that requests (which is the base URL http://www.xxxxxx.com/ ) With regards to the HTTP responses the reverse proxy is ignoring them (override-expire ignore-reload ignore-no-cache) and Squid decides what to cache, what not to cache, etc. by the rules based on acl, refresh_pattern.... -- View this message in context: http://squid-web-proxy-cache.1019090.n4.nabble.com/Create-acl-based-on-Accept-Language-tp4661764p4661773.html Sent from the Squid - Users mailing list archive at Nabble.com.