Re: [TEST/REVIEW] Netflow implementation

From: Gleb Smirnoff (glebius_at_freebsd.org)
Date: 09/10/04

  • Next message: Vlad GALU: "Re: [TEST/REVIEW] Netflow implementation" 
    Date: Fri, 10 Sep 2004 09:36:19 +0400
To: Vlad GALU <dudu@diaspar.rdsnet.ro>

    On Thu, Sep 09, 2004 at 11:41:26PM +0300, Vlad GALU wrote:
    V> This made me raise my eyebrow. I wrote a small tool that we use in
    V> production at RDS: http://freshmeat.net/projects/glflow. The way I
    V> designed it, it is supposed to clean up the flow tree once in a while
    V> and remove 'old' flows (that haven't had any packet matching them in the
    V> last X seconds). The problem is that I currently have about 400-500k
    V> active flows on a 700Mbps link. Every 10 seconds the software removes
    V> about 100-200k of them in no more than 0.2-0.3 seconds. Of course, I
    V> couldn't possibly send them over a socket somewhere else at that speed,,
    V> and chose to open a tempfile, mmap() it, write the expired flows to the
    V> buffer. When the buffer exceeds a programatically chosen number of
    V> packets, it is msync()-ed, munmap()-ed and a new file is open.

    If you remove 100-200k of flows every 10 seconds, this means you have 10 - 20
    kpps of worm traffic or more! Impressive.
    200k flows expired in 10 seconds is 666 export dgrams per second, 7 times
    more than I usually have in my testbed. Not sure, that ng_netflow + ng_ksocket
    can do this.

    Do you collect all this traffic on a single router?

    V> Do you accidentally have a better storage model ? I've been trying to
    V> dump these binary files to SQL but for a 42 meg binary log the necessary
    V> SQL storage went to about 150 megs, which is a bit over reasonable,
    V> considering the fact that the software dumps a binary file every 5 to 10
    V> seconds.

    Dumping to SQL is a bad idea. I have tried it, too :)

    V> P.S. I haven't yet tried to aggregate the flows between reading them
    V> from the binary file and inserting the data into SQL. I thought it would
    V> take too much time to be able to keep up with the newly created dumps.

    This is how many people do. However I don't know details. 

    -- 
Totus tuus, Glebius.
GLEBIUS-RIPN GLEB-RIPE
_______________________________________________
freebsd-net@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-net
To unsubscribe, send any mail to "freebsd-net-unsubscribe@freebsd.org"
  • Next message: Vlad GALU: "Re: [TEST/REVIEW] Netflow implementation"

    Relevant Pages

    • Re: [TEST/REVIEW] Netflow implementation
      ... > Full netflow dumps are stored on disk for about 2-3 months, ... active flows on a 700Mbps link. ... packets, it is msync-ed, munmap-ed and a new file is open. ... from the binary file and inserting the data into SQL. ...
      (freebsd-net)
    • Re: [TEST/REVIEW] Netflow implementation
      ... > V> dump these binary files to SQL but for a 42 meg binary log the ... > V> necessary SQL storage went to about 150 megs, ... I haven't yet tried to aggregate the flows between reading ... > V> newly created dumps. ...
      (freebsd-net)