Local caching proxy server

Has anyone considered setting up a local proxy server, vm or whatever, implementation isn't too critical... to cache certain requests?  If you had a 100gb cache, that's a potentially huge speed-up.

Certain sites just would not work with the caching proxy due to needing a recent copy of the page, like 4chan.  But, perhaps you could cache certain youtube videos / porn links that you tend to look at often.  Instant loading ftw.  Or, pre-emptive caching...  load a certain page, and it will perform background requests to all links on that page in order to speed up browsing.  Pre-emptively cache a page and all immediate subpages that your household might look at daily, for example, news sites.  Could be handy for wikipedia as well, because chances are you might want to look at that article again.

Have every cached paged served up with a <div> on the top saying something like, "Cached on xx:xx:xx:xx, renew?" clicking on it would resend the request.

With the advent of huge amounts of storage space in your typical desktop (a new desktop these days could easily have 2TB), 100gb cache is not so bad.  Bandwidth is typically not an issue.  Use the LRU algorithm to detect which item to delete from the cache.

Of course, the only concern is that the increased web browsing traffic fucks up gaming.  For this... 2 separate lines, one dedicated to real-time tasks (gaming, VOIP, email, etc), and the other dedicated to background network tasks (file transfer, web browsing)

If you're not rich (like most of us), have the proxy detect when someone is gaming and disable any pre-fetching.

What do you guys think?

Holy shit, has /prog/ really come to this?

so basically, what you're saying is "what if we set firefox's history cache to 100GB".

100gb cache
（　≖‿≖）

>>3
No.
An actual proxy server such that your local entire network can benefit, not just 1 computer...
>>2
Didn't want to put it in /technology/ or whatever; I looked at the topics, saw "uggz", and decided to come here.  You'd rather argue about SICP and other stupid shit?
>>4
I appreciate your comment the most

Doesn't having a huge cache kind of defeat the purpose or am I just retarded(a very real possibility)?

Back to /shoes/, please

>>1
>>5
There are a shitload of these. Polipo, squid, privoxy, etc. There are a lot more optimizations that can be done besides caching, such as ad stripping, pipelining, etc.

>>1
I must say, not a bad idea.

I think it would benefit most from only caching non html though.  Images, sounds, videos, all the bulk of the internet.  The only things that change on news sites, blogs, etc. is text.  So if you only fetch updated text and cache the other content not only will it be wicked fast, but updated too.  When you request headers for html pages you can see the date last updated too, so in most cases you don't need to fetch any content.

>>8
Listen to this man, he speaks the truth.
As a starter consider a sqiud proxy.
To add to >>8-chan's mess: an haproxy is fine too

>>9
Then you lose the capacity for some offline browsing.  I suppose you need site specific roles; what you described with the media caching only would be good for a household of 4chan users.

http://en.wikipedia.org/wiki/Squid_%28software%29
Get the fuck out.

>>1
Dude your browser already does this. Go to options and drag the cache size slider way over to the right. Done.

Also, you may not know this but HTTP/HTML already have a lot of functionality to let the browser determine whether a resource has changed before needing to download it (HEAD requests, expiration headers, etc). This is why you don't need some div to manually refresh pages; it happens automatically, and shift+refresh exists in case it breaks. Just let the tools you already have do their job.

And if you really want to speed up browsing, honestly the best thing you can do is install adblock. Don't fuck around with a proxy; in fact they usually make things slower because of all the useless redundant parsing they do of all your downloaded shit.

Why is it called Squid? Because it's fast and slippery?

>>13
Sensible advice? On my /prog/?

>>15
This may surprise you, but yes, I like to give sensible advice on occasion. It's just that most queries to /prog/ are so stupid (quick search in Google returns something) that they're not even worth doing anything useful.