|
| quote: | Originally posted by Durafei
I wonder what pipelining means in Firefoxes case. The HTML itself is received over one TCP connection after a GET request. They can't really speed that up. Do they just open multiple connections to get multiple images from the page ? I thought they would do that by default... |
It's just the bundling of requests within a given tcp packet. Instead of sending 1 then waiting for a response. You send say 8 at once, then sit and wait for the response from those 8. You are correct in that the original GET request for the HTML isn't going to go any faster. It's not a pipelined request. After the browser determines the web servers capabilities (most should support it, with HTTP/1.1), the rest of the requests can all be 'pipelined'. Everything takes place within a single connection, that has been true before enabling pipelining, the difference is the requests were sequential for the remaining content, rather then parallel (in a way). The HTML has say 8 images to retrieve, instead of requesting each one in sequence, not proceeding to the next one until a response is received from the previous request, you simply send the request for all 8 at once in a single packet, then wait for the response from all 8.
Read the FAQ from Mozilla, http://www.mozilla.org/projects/net...lining-faq.html
|
