Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revision Previous revision
Next revision
Previous revision
linux:iptables:l7patch [2007/06/09 17:31]
a
linux:iptables:l7patch [2009/05/25 00:35] (current)
Line 48: Line 48:
 You will have now every p2p packet marked, but a dramtic reduce of udp You will have now every p2p packet marked, but a dramtic reduce of udp
 mismatches. mismatches.
 +
  
 ===== Yet another way to do it .. ===== ===== Yet another way to do it .. =====
Line 81: Line 82:
    -fPIC -c  libipt_ipp2p.c    -fPIC -c  libipt_ipp2p.c
    ld -shared -o libipt_ipp2p.so libipt_ipp2p.o    ld -shared -o libipt_ipp2p.so libipt_ipp2p.o
 +
 +==== Which CPU for heavy traffic with much filtering/shaping ====
 +<code>
 +> Subject: Re: [LARTC] Which CPU for heavy traffic with much
 +> filtering/shaping?
 +> Date: Mon, 19 Nov 2007 17:40:34 +0100
 +
 +> >Hi
 +
 +> Hi
 +
 +> >I have a router with a large number of iptables rules and some
 +> >extensive traffic shaping (HTB + RED + ... ) + conntrack.
 +
 +> Performance boost tips:
 +
 +> - Use "set" module instead of sequential iptables rules. It can lower
 +> cpu usage.
 +
 +> - Use hashing filters for shaping if you're using many u32 filters.
 +
 +> - configure conntrack to use bigger hashsize for better performance;
 +> i'm passing following parameter to kernel in grub to achieve this:
 +> ip_conntrack.hashsize=1048575 
 +
 +> - configure routecache to use bigger to use more memory for better
 +> performance; i'm passing following parameter to kernel in grub to
 +> achieve this: rhash_entries=2400000
 +
 +> >1. What processors should I be looking for in order to achieve the
 +> >best routing throughput on a linux router?
 +
 +> I've had good experiences with P4 (with and without HT), Athlon64, Xeon
 +> [dempsey], Xeon [woodcrest]. The last one is the best choice because of
 +> the large cache and architecture. I think you can use Core 2 Duo too
 +> if you want to save some money.
 +
 +> >2. Is it true that multicore processors will not help much in this 
 +> >situation?
 +
 +> Not true. In your setup with two nics with same load you can easily use
 +> two cores. You can assign each nic to different core by the means of
 +> smp_affinity setting in /proc/irq/... or by using irqbalance daemon.
 +</code>
 +
 +==== L7filtering + CentOS ====
 +   * [[http://www.howtoforge.com/how-to-set-up-a-linux-layer-7-packet-classifier-on-centos5.1]]
 +
linux/iptables/l7patch.1181403081.txt.gz ยท Last modified: 2009/05/25 00:34 (external edit)
CC Attribution-Share Alike 4.0 International
Driven by DokuWiki Recent changes RSS feed Valid CSS Valid XHTML 1.0 ipv6 ready