LinuxQuestions.org
Register a domain and help support LQ
Go Back   LinuxQuestions.org > Forums > Linux Forums > Linux - Hardware
User Name
Password
Linux - Hardware This forum is for Hardware issues.
Having trouble installing a piece of hardware? Want to know if that peripheral is compatible with Linux?

Notices


Reply
  Search this Thread
Old 02-12-2014, 12:18 PM   #1
ccc
Member
 
Registered: Oct 2002
Posts: 94

Rep: Reputation: 0
Minimal hardware requirement for Squid Proxy


hi

I'm planning to setup Squid Caching Proxy running on Linux for about 800 clients?
Can anyone help me pls with hardware requirements, to get possibly best performance?

What do you thing about this link?

http://beginlinux.com/blog/2008/11/h...nts-for-squid/
>>>
Hardware Requirements for Squid


by Mike on November 7, 2008 1 comment

in Proxy Server

The hardware requirements are not as large as you would think. The most important aspect to consider is the RAM that is available for Squid. RAM is important because each object in the cache requires a small amount of memory. Generally, 32 MB of RAM are required for every GB of disk space. If you run out of memory there will be a significant reduction in speed.

The other major consideration for Squid is disks. The faster the disk read and write the faster Squid will operate. Usually it is a good idea to consider SCSI for disks on a proxy server just because of speed. The other advantage that SCSI has is that it can access 7 different drives allowing for multiple reads and writes without a slowdown in access. If you are using ATA drives and have multiple drives on one channel you will find the system has to wait as it can only access one drive at a time. However, SATA drives or even some ATA drives are increasing in speeds and are much cheaper.
There are a number of variables that impact the speed of Squid and the hardware that is required. One variable is object size. The larger the object, the more memory is required per object so this may increase memory requirements. The second variable is the number of users that are on the system concurrently. This is a large variable in that the difference between 5 users and 105 users is considerable. The point is, plan for growth and estimate high for concurrent users so you do not need to come back later and upgrade.
<<<

Last edited by ccc; 02-13-2014 at 11:14 AM.
 
Old 02-12-2014, 12:25 PM   #2
TB0ne
Guru
 
Registered: Jul 2003
Location: Birmingham, Alabama
Distribution: SuSE, RedHat, Slack,CentOS
Posts: 15,990

Rep: Reputation: 2972Reputation: 2972Reputation: 2972Reputation: 2972Reputation: 2972Reputation: 2972Reputation: 2972Reputation: 2972Reputation: 2972Reputation: 2972Reputation: 2972
Quote:
Originally Posted by ccc View Post
hi
I'm planning to setup Squid Caching Proxy running on Linux for about 800 clients? Can anyone help me pls with hardware requirements, to get possibly best performance?
Much like your previous question:
http://www.linuxquestions.org/questi...ce-4175483956/

...and the answer doesn't change. You leave out too many variables for anyone to give a meaningful answer. And again, please don't use text-speak.
Quote:
What do you thing about this link?
http://beginlinux.com/blog/2008/11/h...nts-for-squid/
>>>
Hardware Requirements for Squid


by Mike on November 7, 2008 1 comment

in Proxy Server

The hardware requirements are not as large as you would think. The most important aspect to consider is the RAM that is available for Squid. RAM is important because each object in the cache requires a small amount of memory. Generally, 32 MB of RAM are required for every GB of disk space. If you run out of memory there will be a significant reduction in speed.

The other major consideration for Squid is disks. The faster the disk read and write the faster Squid will operate. Usually it is a good idea to consider SCSI for disks on a proxy server just because of speed. The other advantage that SCSI has is that it can access 7 different drives allowing for multiple reads and writes without a slowdown in access. If you are using ATA drives and have multiple drives on one channel you will find the system has to wait as it can only access one drive at a time. However, SATA drives or even some ATA drives are increasing in speeds and are much cheaper.
There are a number of variables that impact the speed of Squid and the hardware that is required. One variable is object size. The larger the object, the more memory is required per object so this may increase memory requirements. The second variable is the number of users that are on the system concurrently. This is a large variable in that the difference between 5 users and 105 users is considerable. The point is, plan for growth and estimate high for concurrent users so you do not need to come back later and upgrade.
<<<
What about it? Makes sense, but this also applies to pretty much EVERY piece of software...give it fast drives and lots of memory, and you will very obviously have it run faster.
 
Old 02-12-2014, 02:20 PM   #3
ccc
Member
 
Registered: Oct 2002
Posts: 94

Original Poster
Rep: Reputation: 0
[QUOTE=TB0ne;5116334]Much like your previous question:
http://www.linuxquestions.org/questi...ce-4175483956/

That was an other project, half a year ago.
The question was about config file and this proxy is already running.
 
Old 02-12-2014, 02:38 PM   #4
TB0ne
Guru
 
Registered: Jul 2003
Location: Birmingham, Alabama
Distribution: SuSE, RedHat, Slack,CentOS
Posts: 15,990

Rep: Reputation: 2972Reputation: 2972Reputation: 2972Reputation: 2972Reputation: 2972Reputation: 2972Reputation: 2972Reputation: 2972Reputation: 2972Reputation: 2972Reputation: 2972
Quote:
Originally Posted by ccc View Post
Quote:
Originally Posted by TB0ne View Post
That was an other project, half a year ago. The question was about config file and this proxy is already running.
And again, the answer remains the same. You have not provided enough details, and asked a vague question. You can't get a meaningful answer without having the details, and providing context.

To explain it to you again, 800 users is meaningless, unless
  • You say how many of the 800 you expect to be connected at any one time
  • How those 800 users authenticate
  • The rules/ACL's you have in place
  • What kind of caching you're doing
  • What kinds of pages they're going to be accessing
  • ..and MANY other criteria, like how many users are going to be added to the box/growth rate, and over what period of time, etc.
You're looking for someone to give you server specs, without knowing how it's going to be used...and no one can. Buy the fastest server you can afford, with as much memory as it will hold, and the fastest disks you can find. That's the most accurate recommendation anyone can give you, based on what you say. After being registered here for TWELVE YEARS now, these things should be obvious. They're not Linux specific...asking someone "How can I build a Windows server for 800 users?" will get a similar response, so even if you haven't worked with Linux for 12 years, the criteria remain the same.
 
Old 02-12-2014, 03:04 PM   #5
ccc
Member
 
Registered: Oct 2002
Posts: 94

Original Poster
Rep: Reputation: 0
Sorry for troubles:

[*]You say how many of the 800 you expect to be connected at any one time
400-500 in HQ and branches over MLPS
[*]How those 800 users authenticate
AD policy authentication
[*]The rules/ACL's you have in place
acl to allow different WAN networks
[*]What kind of caching you're doing
Normal, (Regular/Caching) Proxy which listens on a separate port (e.g. 3128) and the clients (browsers) are configured to send requests for connectivity to that port. So the proxy server receives the request, fetches the content and stores a copy for future use. So next time when another client requests for the same webpage the proxy server just replies to the request with the content in its cache thus improving the overall request-reply speed
[*]What kinds of pages they're going to be accessing
DansGuardian
[*]..and MANY other criteria, like how many users are going to be added to the box/growth rate, and over what period of
only 800 users for 2-3 next years.

Additional features:
squidGuard and clamAV running behind on the same machine.

Last edited by ccc; 02-12-2014 at 03:06 PM.
 
Old 02-12-2014, 04:41 PM   #6
TB0ne
Guru
 
Registered: Jul 2003
Location: Birmingham, Alabama
Distribution: SuSE, RedHat, Slack,CentOS
Posts: 15,990

Rep: Reputation: 2972Reputation: 2972Reputation: 2972Reputation: 2972Reputation: 2972Reputation: 2972Reputation: 2972Reputation: 2972Reputation: 2972Reputation: 2972Reputation: 2972
Quote:
Originally Posted by ccc View Post
Sorry for troubles:[*]You say how many of the 800 you expect to be connected at any one time
400-500 in HQ and branches over MLPS
[*]How those 800 users authenticate
AD policy authentication
[*]The rules/ACL's you have in place
acl to allow different WAN networks
[*]What kind of caching you're doing
Normal, (Regular/Caching) Proxy which listens on a separate port (e.g. 3128) and the clients (browsers) are configured to send requests for connectivity to that port. So the proxy server receives the request, fetches the content and stores a copy for future use. So next time when another client requests for the same webpage the proxy server just replies to the request with the content in its cache thus improving the overall request-reply speed
[*]What kinds of pages they're going to be accessing
DansGuardian
[*]..and MANY other criteria, like how many users are going to be added to the box/growth rate, and over what period of
only 800 users for 2-3 next years.

Additional features:
squidGuard and clamAV running behind on the same machine.
Ok, that's SOME of the details, but you're STILL lacking a good bit of detail. Again, as an administrator who has been here for twelve years, you should know the kinds of details you need to look at to size ANY server.

Again, to be VERY CLEAR:
buy the fastest server you can afford, put as much memory in it as you can, and get the fastest disks you can find. Nothing much else we can tell you.
 
Old 02-13-2014, 11:13 AM   #7
ccc
Member
 
Registered: Oct 2002
Posts: 94

Original Poster
Rep: Reputation: 0
Quote:
Originally Posted by TB0ne View Post
Again, to be VERY CLEAR:
buy the fastest server you can afford, put as much memory in it as you can, and get the fastest disks you can find. Nothing much else we can tell you.
Perhaps has anyone squid running for so many users and would like tell me what Hardware will be used.
At least I should know minimal hardware requirement.
My second problem is to choose, if run on own hardware, or as guest VM in virtual environment.
 
Old 02-13-2014, 11:48 AM   #8
TB0ne
Guru
 
Registered: Jul 2003
Location: Birmingham, Alabama
Distribution: SuSE, RedHat, Slack,CentOS
Posts: 15,990

Rep: Reputation: 2972Reputation: 2972Reputation: 2972Reputation: 2972Reputation: 2972Reputation: 2972Reputation: 2972Reputation: 2972Reputation: 2972Reputation: 2972Reputation: 2972
Quote:
Originally Posted by ccc View Post
Perhaps has anyone squid running for so many users and would like tell me what Hardware will be used.
...which will, AGAIN, not apply to your situation, since each environment is different. AGAIN, you are leaving out far too many details, and there really isn't room/time on a public forum for folks to gather everything. If you're having trouble, then hire a consultant to do this for you.
Quote:
At least I should know minimal hardware requirement.
There really is no 'minimum', and there won't be no matter how many times you ask. You can run it on a $50 netbook...it will be slow, but it will work. So, by that definition (the software FUNCTIONS for what you need), that's the 'minimum requirement'.
Quote:
My second problem is to choose, if run on own hardware, or as guest VM in virtual environment.
That's still not a problem at all...pick one. You can run it on VM or real hardware, and as long as the resources are the same, it'll function the same.

PLEASE PAY ATTENTION: there are TOO MANY UNKNOWNS for anyone here to tell you anything meaningful. No matter how many times you ask. Saying "best performance" is meaningless, since 'best performance' for me might not be what YOU want, or vice-versa. If you don't understand the very simple answer you're getting, then you need to hire someone to do this for you; you have been an administrator for TWELVE YEARS, and should know the basics of how to size a server by now. AGAIN: buy the fastest server you can afford, with as much memory as it can hold, with the fastest disks you can put in it. Whether that's a VM server or not is immaterial.

A brief Google search turns up MANY items on how to build/size a Squid server...all of them point to different things. A Pentium 4 has handled 500 users, while an i5 has trouble. It depends on your environment:
http://www.deckle.co.uk/squid-users-...ing-squid.html

Last edited by TB0ne; 02-13-2014 at 01:12 PM.
 
Old 02-13-2014, 03:48 PM   #9
ccc
Member
 
Registered: Oct 2002
Posts: 94

Original Poster
Rep: Reputation: 0
Anyway, I don't have any questions.

I will start with this calculation:
10-15MB per 1 GB of cache_dir.
each live connection consumes about ~60-70 KB of RAM.

I will setup squid on virtual machine, so I can later resize hard disk, add a second CPU or resize RAM.
 
  


Reply

Tags
browser, caching, proxy, squid


Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is Off
HTML code is Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
web squid proxy behind a hardware firewall NAT device? luvlinux2009 Linux - Networking 1 02-22-2010 08:12 PM
PXE server hardware requirements (probably NFS server hardware requirements) Ipolit Linux - Networking 0 07-27-2009 01:33 AM
Hardware recommendations to run squid proxy 24x7 dennisbrain57 Linux - Newbie 1 06-24-2009 05:55 PM
configure squid proxy with microsoft proxy as a parent proxy nintykola Linux - Software 1 08-28-2007 01:38 AM


All times are GMT -5. The time now is 12:06 AM.

Main Menu
Advertisement
My LQ
Write for LQ
LinuxQuestions.org is looking for people interested in writing Editorials, Articles, Reviews, and more. If you'd like to contribute content, let us know.
Main Menu
Syndicate
RSS1  Latest Threads
RSS1  LQ News
Twitter: @linuxquestions
Facebook: linuxquestions Google+: linuxquestions
Open Source Consulting | Domain Registration