When people try to use web page pre-fetching on their local network (on their side of the internet pipe/DSL/etc), often this result in lower performance. But I have a case where I would like to try it but am having a hard time finding good documentation etc.

Here is my use case:

1. My internet connection to my linode is great! I am fortunate that path from Cincinnati to the Atlanta nodes is short and fast.
2. The linode itself has much MUCH faster access to the internet than my cable model.

As a result, pre-fetch would have no adverse impact on my SOHO network performance, and I have gobs of bandwidth etc. on my node to use for this.

The problem is when I install squid3 and squid-prefetch, and watch the squid log I don't see any pre-fetching happening ! Also, there seems to be little to no documentation on pre-fetch.

Is anyone doing this?
Anyone using an alternative strategy to accelerate their SOHO web experience?

Thanks, Erik