Search squid archive

Load-balanced Squid, "always_direct", and Cached Content

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

 



If "always_direct" is used for systems in a specific domain, will a Squid 
proxy query its peers for cached content?

   (1)	WPAD is defined in the local DNS zone to support the automatic
	detection of web proxies.  Assuming that automatic detection is
	enabled, Linux- and older Windows- based will locate and use the
	proxy.pac/wpad.dat configuration file; however, WindowsXP-based
	systems will not.
   (2)	DHCP is configured to pass Option 252 to pass the URL for the web
	proxy automatic configuration script.  Assuming that automatic
	detection is enabled, WindowsXP-based systems will only use the
	proxy.pac/wpad.dat file when configured for DHCP and Option 252
	is defined to the URL. 
   (3)	The same proxy.pac/wpad.dat file that defines a simple hashing
	algorithm to balance the load between two Squid proxy servers.

When a Mozilla-based browser (FireFox, Mozilla, Netscape, etc.) is used, 
load-balancing works as expected.  When Internet Explorer is used, the 
load-balancing fails.  If a specific web server has never been accessed, 
the hashing alforithm is used to select the web proxy.  However, once a 
proxy selection has been made, Internet Explorer will use that proxy for 
all accesses to that specific web server.

The load-balanced proxy servers have a common parent that sits on the 
network security perimeter to provide access to Internet content.  When 
content is being retrieved from the Internet it doesn't matter whether 
one Windows system selected proxy A and another selected proxy B.  If the 
content is cacheable, it will be retrieved from the parent's cache.

The parent proxy can access both external Internet content and internal 
corporate content over our wide area network.  At present, "always_direct" 
is only used on the load-balanced proxy serversto access web servers in 
our local facility.  I would like to use "always_direct" on the proxies
for all internal corporate web content.

If all browsers used the hashing algorithm for all requests, there would 
be no question about using "always_direct" as the browser would select the 
proxy that would have the content cached.  The question is does the proxy 
query its peers before applying the "always_direct" rule?  If not, does 
this behaviour change when using NetDB?

Merton Campbell Crockett

 
-- 
BEGIN:				vcard
VERSION:			3.0
FN:				Merton Campbell Crockett
ORG:				General Dynamics Advanced Information Systems;
				Intelligence and Exploitation Systems
N:				Crockett;Merton;Campbell
EMAIL;TYPE=internet:		mcc@xxxxxxxxxxxxxxx
TEL;TYPE=work,voice,msg,pref:	+1(805)497-5045
TEL;TYPE=work,fax:		+1(805)497-5050
TEL;TYPE=cell,voice,msg:	+1(805)377-6762
END:				vcard

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

  Powered by Linux