Search squid archive

Recommendations for a fast cache with Squid+WCCP2

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

 



Hi Cache-ists,

I hate to ask this question again, as I faintly can recall something 
like it being asked, but allow me to ask just to be sure.

Now that I know how to get COSS working, I'd like to build a "good"
cache using Squid 2.6STABLE4 (well, STABLE5 if it gets to the ports
soon).

I intend to use WCCP2, which I am not familiar with, but thanks to 
Adrian, I have seen the following example configurations:

http://wiki.squid-cache.org/ConfigExamples/MultiplePortsWithWccp2?highlight=%28configex%29
http://wiki.squid-cache.org/ConfigExamples/NatAndWccp2?highlight=%28configex%29

I am going to share with our network engineers who man the
Cisco routers. I have never touched this WCCP stuff so I have absolutely 
no idea about it, except for what I have seen being discussed.

I have also read a very nice article, though Linux-centric, about 
building a high performance cache using squid:

http://www.devshed.com/c/a/Administration/High-Performance-Web-Caching-With-Squid/2/

... but I live on the other side ... the BSD side ;)


Basically, we are going to try and save bandwidth by using the cache. 
It's important that it performs, so that customers don't start 
quarrelling with us about slow speeds.

We have a few thousand clients on our network (196.200.32.0/20 and
62.8.64.0/19). Not all subnets are used though, but the intention
is to have all the active subnets on the cache. Generally, my intention 
is build a large, reliable, fast cache.

I have a few questions:

1. What is the difference between using transparent interception, for
   example via a IPNAT, and using WCCP?

I am familiar with transparent proxy and I know it doesn't play well with https.
How about a WCCP implementation?

2. Would WCCP method be the recommended path? Why? (this Q seems 
   redundant, yes?)

3. What hardware recommendations would you give for such a large 
   network?? CPU, RAM, Disks?
   Would a RAID array do? Or would caches spread over several single
   disks do better?

4. Would COSS be a good choice, or should I stick to aufs for this one?


Thank you in advance.



-Wash

http://www.netmeister.org/news/learn2quote.html

DISCLAIMER: See http://www.wananchi.com/bms/terms.php

--
+======================================================================+
    |\      _,,,---,,_     | Odhiambo Washington    <wash@xxxxxxxxxxxx>
Zzz /,`.-'`'    -.  ;-;;,_ | Wananchi Online Ltd.   www.wananchi.com
   |,4-  ) )-,_. ,\ (  `'-'| Tel: +254 20 313985-9  +254 20 313922
  '---''(_/--'  `-'\_)     | GSM: +254 722 743223   +254 733 744121
+======================================================================+

Gnagloot, n.:
	A person who leaves all his ski passes on his jacket just to
impress people.
		-- Rich Hall, "Sniglets"

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

  Powered by Linux