From: Casey Scott <casey@phantombsd.org>
To: Randy Grimshaw <rgrimsha@syr.edu>
Cc: netfilter@lists.netfilter.org
Subject: Re: Throttling NAT interface
Date: Wed, 5 Apr 2006 11:17:41 -0700 (PDT) [thread overview]
Message-ID: <19459423.301144261061585.JavaMail.root@tomcat.phantombsd.org> (raw)
Do you think that REJECTing such packets will cause the sending
device to reduce its transmitting speed?
Casey
----- Original Message -----
From: Randy Grimshaw <rgrimsha@syr.edu>
To: casey@phantombsd.org
Cc: netfilter@lists.netfilter.org
Sent: Wednesday, April 5, 2006 11:11:16 AM GMT-0800
Subject: Re: Throttling NAT interface
I hope this isn't evasive, but I am still learning this stuff too.
The packets that don't match - i.e. exceed the threshold - will continue down the chain.
You can do a lot of different things with those packets as you need. You can probably match the behavior of tc if that is what you want.
<><Randy
<><Randall Grimshaw
Room 203 Machinery Hall
Syracuse University
Syracuse, NY 13244
315-443-5779
rgrimsha@syr.edu
>>> Casey Scott <casey@phantombsd.org> 4/5/2006 1:35 PM >>>
Thanks for the response. Will using iptables in that
manner cause the firewall to just drop the packets that
arrived after the limit has been reached? I don't want
to have to drop packets that have already used up
bandwidth getting to the machine, and making matters
worse by requiring the packets to be sent again.
Thanks,
Casey
----- Original Message -----
From: Randy Grimshaw <rgrimsha@syr.edu>
To: netfilter@lists.netfilter.org, casey@phantombsd.org
Sent: Wednesday, April 5, 2006 10:27:48 AM GMT-0800
Subject: Re: Throttling NAT interface
Our throttle is really in the other direction but the
idea may work for you.
iptables [traffic pattern spec] -m limit --limit
40/second --limit-burst 60 -j LEVEL2
LEVEL2 is a user defined chain that specifies other
restrictions or in your case NAT translations.
The point is that you can match the traffic patterns
that you need before applying the limit match.
hope this is helpful.
<><Randy
<><Randall Grimshaw
Room 203 Machinery Hall
Syracuse University
Syracuse, NY 13244
315-443-5779
rgrimsha@syr.edu
>>> Casey Scott <casey@phantombsd.org> 4/5/2006 12:44 PM >>>
The gist of what I need to do is restrict the rate of
off-network traffic coming in through a host. The host
is providing basic NAT to an internal network. I have
gotten pretty close to what I need to do with iptables
and tc. The problem is that when an interface is throttle
with tc, the source of the traffic doesn't matter. I don't
want to throttle local traffic, just traffic coming through
the machine from a WAN. The next step was to add another
NIC to the machine. Something like this:
eth0: <local IP1>
eth1: <WAN IP>
eth2: <local IP2>
The problem is that even if traffic destined to be routed
off network comes into eth2, which is throttle via tc, the
return traffic comes back through eth0. Since tc (tbf filter)
just controls the transmitting of an interface, I need to
force the NAT traffic to use eth2. Traffic that is meant to
stay local can use eth0. Is possible to do something like this?
How can I this host to only eth2 for NAT even though both eth0
and eth2 are in the same network? eth0 is not throttled, which
is why local traffic needs to use it.
TIA,
Casey
next reply other threads:[~2006-04-05 18:17 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2006-04-05 18:17 Casey Scott [this message]
-- strict thread matches above, loose matches on Subject: below --
2006-04-05 18:11 Throttling NAT interface Randy Grimshaw
2006-04-05 17:35 Casey Scott
2006-04-05 17:27 Randy Grimshaw
2006-04-05 16:44 Casey Scott
2006-04-05 15:22 Casey Scott
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=19459423.301144261061585.JavaMail.root@tomcat.phantombsd.org \
--to=casey@phantombsd.org \
--cc=netfilter@lists.netfilter.org \
--cc=rgrimsha@syr.edu \
/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.