Half a croissant, on a plate, with a sign in front of it saying '50c'
h a l f b a k e r y
Reformatted to fit your screen.

idea: add, search, annotate, link, view, overview, recent, by name, random

meta: news, help, about, links, report a problem

account: browse anonymously, or get an account and write.



distributed proxy

every browser becomes a proxy for all computers on local network.
  [vote for,

I like proxys but I don't like the fact I have to set a proxy up - terrible on my laptop while I move around. I wondered if there might be away of doing a peer to peer proxy. In essence a browser (or local machine proxy ) when it wanted to get say an image from the internet would sent out a multihost UDP ping packet with the hash of the URL. Every other machine running the browser/local proxy on the local network ( anything from XXX.XXX.XXX.001 to XXX.XXX.XXX.255 ) would then look up the hash in its own internal cache and if it finds it would then send a YES UDP packet back.

The browsing machine takes the first YES UDP packet it gets back. It then does a TCP connection to the machine that said YES over port 80 and down loads the cached copy of the image FROM the YES machine. If no local machine says yes within a small time then the browsing machine goes over the internet as normal.

Later if another machine wants the same image URL then the browsing machine then acts as the YES machine and you server the image from your cache.

This first past the post system automatically load balances over the machines on the local network. If a number machines have the same image then the first to respond is going to be the fastest/least loaded machine.

Usual cacheing proxy rules apply ( don't cache images with /cgi/ or .php in url and so on).

points for extra credit.

1- adverts might get mangled to be replaced with adverts for stuff you already own but haven't used in a bit. 2- a serve simply becomes a machine that listens to requests and then down loads the HTML. the HTML is scanned for links and the links are pre-loaded. Thus when a browsing machine calls the server has simply pre-cached the target pages. 3-If someone walks in you your local cyber cafe with a hardrive full of cached goodies then they are acting as a walking connection to the internet.

inventorist, Apr 15 2007


Voice, Apr 16 2007

       From my (limited) understanding, this is kind of how the internet works already. A single page of data might be transmitted via multiple routes (via multiple proxies/routers), with packets arriving at their intended destination asynchronously - whether this distributed/asynchronous bit happens on the inside or the outside of your subnet isn't going to make a great deal of difference, is it?
zen_tom, Apr 16 2007

       The problem with this is that every browser has its own cache, so the proxy would have to work with each seperately. First thing we need is a standard http layer that all browsers can utilize and gain a shared cache.
ironfroggy, Apr 17 2007

       Good idea! Some Cons: - UDP works bad on high load networks. - Security is low. Due spreading unencrypted (http) private data over all local network computers. So, any user can spy for any :)
Alexey Vdovin, Jan 13 2009

       This is basically Squid meets Tor. Since all the users mirrors would become mostly homogeneous eventually, the principles of Bittorrent could be applied, too. I have no idea where this sort of system would end up being useful, but the idea is interesting so [+].
Spacecoyote, Jan 13 2009


back: main index

business  computer  culture  fashion  food  halfbakery  home  other  product  public  science  sport  vehicle