Re: [apops] using NTP/etc to measure n/w latency?

  • To: "Robert B. Cohen" <bcohen at bway dot net>
  • Subject: Re: [apops] using NTP/etc to measure n/w latency?
  • From: "Henk Uijterwaal (RIPE-NCC)" <henk at ripe dot net>
  • Date: Thu, 26 Oct 2000 17:25:59 +0200 (CEST)
  • Cc: Neale Banks <neale at lowendale dot com dot au>, apops at lists dot apnic dot net, Philip Smith <pfs at cisco dot com>
  • In-reply-to: <39F61C5D.B95EBF65 at bway dot net>
  • Sender: owner-apops@lists.apnic.net
    • Bob, Neale,
      
      > > > Is there any way to measure traffic volumes on specific circuits with
      > > > this, not just delays?
      > >
      > > What/how are you hoping to measure that can't be better/easier measured
      > > with the likes of router interface octet counters etc?
      
      No, we don't measure this.  
      
      We looked at it, but the RIPE community wasn't too enthousiastic about
      both snooping packets from the piece of ethernet just before the border
      router, or having a black box talk directly to the router.
      
      
      > > Ah...... if the RIPE box could use SNMP to measure traffic volumes then
      > > there might be some interesting correlations to be observed.  Is that your
      > > motivation?
      
      
      > YES, it's the SNMP angle! Could the RIPE box use SNMP to measure
      > traffic volumes. That is my motivation. I would like to see the
      > measures of bandwidth volumes directly so we could see exactly how
      > much traffic is moving, but also some correlations with QOS and
      > traffic volumes.
      
      We've always said that we are interested in reporting our measurements via
      SNMP, then integrate them with existing monitoring systems. There is/was
      also some interest in the RIPE TT Working Group, but when we the current
      test-box hosts to provide us with more details on what they would like to
      see, nobody replied.  
      
      I'm also aware that this topic is being discussed in the IETF SNMPconf WG.  
      At the moment, my idea is to wait until this group has come up with a
      draft, then see if we can do an implementation.
      
      The correlations are something one could already do today.  As a first
      approach, record the info from the router, take our analyzed data, and try
      to correlate them off-line. Then, if there is a correlation, one can talk
      to a port on our box and get the data as it is being recorded, then do the
      correlations real-time.
      
      Henk
      
      ------------------------------------------------------------------------------
      Henk Uijterwaal                    Email: henk.uijterwaal at ripe dot net
      RIPE Network Coordination Centre     WWW: http://www.ripe.net/home/henk
      Singel 258                         Phone: +31.20.535-4414,  Fax -4445
      1016 AB Amsterdam                   Home: +31.20.4195305
      The Netherlands                   Mobile: +31.6.55861746  
      ------------------------------------------------------------------------------
      
      A man can take a train and never reach his destination.
                                                     (Kerouac, well before RFC2780).
      
      
      *             APOPS: Asia Pacific Operations Forum              *
      * To unsubscribe: send "unsubscribe" to apops-request at apnic dot net *