Of course its been done. You have a few options. For example, see if your LB environment lets you load balance based on requested URL. Adiran 2008/8/9 Rob Williams <rob.williams@xxxxxxxxx>: > After doing some more research, I wonder if I'm approaching the > problem I want to solve the right way. > > I am creating a large-scale website with a large amount (terabytes) of > static content. What I need is a way to scale access to my static > content. > > I could do this many ways, but an efficient way in my mind would be to > create an array of reverse-proxy cache servers that work together to > represent the static content. > > The architecture would work like this: > > 1) User goes to www.mysite.com > 2) The user's web browser wants to retrieve static.mysite.com/image1.jpg > 3) DNS sends the HTTP request to a load balancer representing port 80 > for static.mysite.com > 4) Load balancer sends request to a random internal reverse proxy > cache server in the cache array > 5) Cache server returns image1.jpg to end user. > > In my example, there would be only one static apache server > representing the terabytes of data. All proxies in the array would > retrieve and cache files from the single web server as needed. > > Now, if no peering protocol is implemented (ICP, HTCP, Cache Digest, > CARP) then eventually all proxies would have the same files waisting > tons of cache memory. Instead, a protocol of some kind should be > implemented so that the cache servers in the array act together like > one big cache. That way as more servers are added to the array, the > system scales to support more cached content. > > Any suggestions thus far? Has anyone done this? > > -Robert > >
