lör 2008-07-05 klockan 20:33 +0200 skrev Dalibor Dukic: > can someone explain me how SQUID handles intercache communication > especially how does it forward HTTP request to parent proxy. By using HTTP. And lightweight ICP for querying neighbour caches to see who may have the requested content. (not for fetching the content, that's done using HTTP) > Is it using one TCP connection with parent proxy or is it opens > separate connection for different HTTP requests? What about pipelined > connections to parent proxy? It's using as many TCP connections as there is concurrent reqests (not connections). For security reasons (and ease of implementation) no pipelining is used, only keep-alive (persistent connections). Regards Henrik
