Search squid archive

Re: Re: HTTP acceleration

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 20/01/2012 11:57 p.m., Rajeev Bansal wrote:
Hi Amos,

   Thanks a lot for addressing the query. I understand, modern browser
can help in HTTP acceleration by opening the multiple TCP connection,
but as per my understanding that will only help in loading the web
page fast as a page may have multiple small objects which browser can
retrieve faster in multiple connections or doing the HTTP pipelining.
But do you think browser can be helpful in accelerating a big file
download (say 50M), because browser will perform the big file download
over the single connection.

"acceleration" and speed is a factor of distance to the data source. The purpose of a cache is to move the data source from some remote server to the close cache machine. Everything you do that decreases proxy caching harms "acceleration".

Squid today can only cache whole objects. So breaking it into small pieces on multiple connections harms the caching.



If I want to accelerate the bigger file download in the proxy (like
Squid) without modifying the client browser, do you think opening the
multiple TCP connections and downloading the different ranges of the
file can help in acceleration.

Why do you ask this now? Your first post was telling us how badly the download manager (which did this) was affecting the network and proxy.

Sending more packets at once does not make the line/wire capacity larger. Just fills it up more and reduces the space available for other things or other clients.

Amos


[Index of Archives]     [Linux Audio Users]     [Samba]     [Big List of Linux Books]     [Linux USB]     [Yosemite News]

  Powered by Linux