All of lore.kernel.org
 help / color / mirror / Atom feed
From: james jones <jame_sj@yahoo.com>
To: lartc@vger.kernel.org
Subject: Re: [LARTC] Kind of curious:
Date: Sat, 08 Mar 2003 03:28:16 +0000	[thread overview]
Message-ID: <marc-lartc-104709418431945@msgid-missing> (raw)
In-Reply-To: <marc-lartc-104702266224362@msgid-missing>

Thanks Martin,

Comments added below:

--- "Martin A. Brown" <mabrown-lartc@securepipe.com> wrote:
> Hello James,
> 
>  : Anyways I was curious if it would be possible to build my own
> linux
>  : router for this purpose and what would I need to install -
> compile - or
>  : patch.
> 
> [ For the record, this is not the only way, but it is one of the
> easiest
>   ways to get into traffic control with linux, and is very well
> supported
>   on this list. ]
> 
>   - kernel with HTB (version 3)  [ 2.4.20+, without patches ]
>   - tc with HTB support:     http://luxik.cdi.cz/~devik/qos/htb/
> 
>  : I am planning on using Gentoo as the base system so I can put
>  : everything I need in the Kernel as I install the system. Problem
> is I
>  : am not sure what to add to the kernel to make this Idea
> possible. Or
>  : even where to find a working example similar to what I want to
> do
>  : with my router.
> 
> See the QoS kernel options.  (Search for "fair queueing" in your
> .config.)
> 
>  : Any pointers will be appreciated.
> 
> As usual, there are two good places to start:
> 
>   http://lartc.org/        LARTC
>   http://www.docum.org/    Stef Coene's site
> 
> There are ample other resources if you have more specific
> questions.
> 
>  : The part 2 portion of the idea would be to have different
> iptables
>  : rules for some of the ip ranges that don't have their bandwidth
>  : limited...
> 
> What exactly do you mean here?  Do you wish to use fwmark as a
> selector
> for traffic control?

Not sure if I mean fwmark. However Stef's docs are on my reading
list. Basically I will want to simulate different types of firewalls
per iprange. Meaning -- the basic firewall config I will allow
anything out and only responses into the "network" and on a more evil
admin type firewall I would like to block various ports from outbound
access.

Any suggestions here would also be appreciated.

>  If so, read some of Stef's docs:
> 
>   http://www.docum.org/
>   http://www.docum.org/stef.coene/qos/docs/filter.html
> 
> Short answer:  Yes, you can use ipchains/iptables to mark packets
> and then
> perform traffic control based on these marks.

Which would be better ipchains or iptables?

> 
>  : What would I need to install/patch to build a bandwidth router?
> The
>  : purpose of the router would be to limit sets of STATIC ip ranges
> to
>  : various up/down speeds for testing purposes.
>  :
> 
> Your example, a possibility.  Modified, with ASCII diagram:
> 
> 
>   selected IPs      bandwidth    class
>   ----------------------------------------
>   192.168.2.4-9      28.8          1:2
>   192.168.2.10-15    56.6k         1:3
>   192.168.2.16-21    128k          1:4
>   192.168.2.22-25    256k          1:5
>   192.168.2.26-30    512k          1:6
> 
>   RB=real bandwidth
>   r = HTB class rate
>   c = HTB class ceil
> 
> Yes you can select ranges of IPs with classifiers (tc filter) and
> restrict
> them to classes with specified bandwidths.
> 
>                 HTB-root-class (1:1) -- - - - - - ->  RBkbit r 
> RBkbit c
>                            |
>     +--------+--------+--------+--------+-------+
> 1:7 |    1:6 |    1:5 |    1:4 |    1:3 |   1:2 |
>     |        |        |        |        |       +-->  28kbit r 
> 28kbit c
>     |        |        |        |        +-- - - - ->  56kbit r 
> 56kbit c
>     |        |        |        +- - - - - - - - - -> 128kbit r
> 128kbit c
>     |        |        +-- - - - - - - - - - - - - -> 256kbit r
> 256kbit c
>     |        +- - - - - - - - - - - - - - - - - - -> 512kbit r
> 512kbit c
>     +-- - - - - - - - - - - - - - - - - - - - - - -> ???kbit r 
> RBkbit c
> 
> Keep in mind, this is but one example of the way you could
> subdivide your
> bandwidth.
> 
> The example above is a description of the classes and children
> classes
> which would allow you to enforce the guidelines you have set out
> above.
> 

Great example you definately got the little mouse back onto the
wheel. 

>  : Again to futher torture the client side applications we test.
> 
> But why would we want to be so vicious to the poor client?  All
> it's
> trying to do is consume every iota of bandwidth....is that so
> wrong?
> The client isn't evil, it's just coded that way....  ;-)


My job is to punish the coders by beating their little client to a
pulp, while maintaining consistant varibles thru out the test. ;-).
The problem is they are starting to get wise to my tricks so I am
having to get more creative... >:-)

James

> -Martin
> 
> -- 
> Martin A. Brown --- SecurePipe, Inc. --- mabrown@securepipe.com

Again Thanks for your help.
_______________________________________________
LARTC mailing list / LARTC@mailman.ds9a.nl
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/

  parent reply	other threads:[~2003-03-08  3:28 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2003-03-07  7:35 [LARTC] Kind of curious: james jones
2003-03-08  2:44 ` Martin A. Brown
2003-03-08  3:28 ` james jones [this message]
2003-03-08  4:07 ` Martin A. Brown
2003-03-08 19:29 ` Stef Coene

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=marc-lartc-104709418431945@msgid-missing \
    --to=jame_sj@yahoo.com \
    --cc=lartc@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.