[LINK] Survey Finds Support for School Filters (USA)

[Danny Yee]

stephen loosley wrote:
> True, school net-access can sometimes indeed be slow, but simple maths shows
> the normal 100 terminals sucking digital-signals over one ISDN line is the problem.
> Thus, from daily experience, and basic common sense I stand by such comments. 
 
I don't really buy the "filtering makes Net access slower" argument
either, at least for schools/organisations with small pipes.  When your
bottleneck is a shared 56k modem or a 64k ISDN line, a little bit of
extra CPU use at either end is going to make SFA of a difference to
the latency.  (Here at the Uni, with a much bigger pipe to the outside
world, pushing everything through a proxy server undoubtedly does
slow things down, but that's done for cost-saving, not for filtering.)

An exception to this, though, would be any kind of filtering that
generates network traffic of its own (through queries to a remote
database somewhere).  That could _really_ slow things down if it were
running on the downstream side of the bottleneck.

Danny.