Search squid archive

Re: anyone describe the model of how Squid manage the memory?

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

 



On 18/08/11 19:26, Raymond Wang wrote:
hi, all:
    I am new to Squid, and I am assigned to learn how Squid manage to
memory, in order to make best use of the Squid.
    there are some problems about the memory management for Squid:

    1, if two files have the same content, such as two Javascript
files,  then how Squid deal with the two files in memory? dose it
treat it as one file and keep the two file name somewhere?
    2, how does the Squid define the level of hot data? and what is the
distribution strategy of hot data like?and How can I affect the
distribution strategy ?


Welcome to the world of caching. :)

Introducing the Squid FAQ, Knowledge Base and How-To collection:
  http://wiki.squid-cache.org/

It's quite big and contains all of your answers, buried somewhere. Enjoy.

Hint:
 http://wiki.squid-cache.org/SquidFaq/SquidMemory

(2) "hot" data to Squid is the set of URLs (a) currently being transferred, plus (b) the N last requested URL objects permitted to stay stored in RAM. cache_mem and maximum_object_size_in_memory control the RAM cache space and object size limits.


(1) Squid deals with URLs and where to find them. That is all. Things like the content of objects at those URLs is completely under the control and responsibility of webmasters authoring the objects. If they have different URLs they are different "URL objects".

(Technical warning)

Notice how I don't say file in any of the above. "file objects" is not really the right idea to be applying if you want to understand the web properly.
 * sometimes one URL object is not a whole file object
 * sometimes one URL object is multiple file objects
* sometimes the URL object can only be described as a "stream of data" or "tunnel". Not related to the concept of "file" in any way.
 * URL object size ranges from zero to infinite (inclusive).
* Sometimes multiple unique URL objects share a URL, the HTTP header meta data then affects potential storage location as well.

On disk cacheable things may look like files. In memory they are structured objects with snippets of HTTP headers and other meta data attached.


Amos
--
Please be using
  Current Stable Squid 2.7.STABLE9 or 3.1.14
  Beta testers wanted for 3.2.0.10


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

  Powered by Linux