On Fri, Mar 14, 2014 at 4:20 PM, Omid Kosari <omidkosari@xxxxxxxxx> wrote: > Really ? Squid has it's own learning mechanism without need human hand ? > Also it can GUESS new urls which it was not aware till now ? Squid doesn't have it's own learning mechanism, it simply does what the helper *you* wrote tells it to do. However, theoretically, it should be possible to compute and store the checksum of the object *after* it's been fetch, and store that in a URL/checksum database. But this requires more than just a StoreID helper, as the helper's role is before the fetch. Something else needs to do the checksum after the fetch. And the database needs to build up its entries first before it becomes useful. The question is whether the additional percentage of hits you get is worth the effort.
