Brett D. Watson bwatson at
Mon Oct 23 21:56:30 UTC 1995

  That was precisely the point of my last note.  I *want* to get involved in 
finding a reliable way to measure network performance and I *want* to be 
proactive at it.  I absolutely hate telling customers that "pinging a cisco is 
not a good indicator..." but like you say "there ain't no other way".  Just as 
ping and traceroute are not good indicators of performance, saying that they 
aren't good indicators is *not* acceptable to customers paying money for your 

 So just for the record, I'm not working for a provider and whining that it's 
not our fault.  I want to find a good way to measure peformance on my network 
and I'd love to be able to publish some results.  Guy Almes contacted me about 
getting involved and that's just what I want to do.


> >  A couple quick points:  the place to go to for this kind
> >of thing should be your direct service provider, who hopefully
> >has some means of communicating with upstream and peer providers
> >when things are going wrong somehow, or at least might be
> >able to give you some additional information.
> Ahem, my success rate at that has been less than stellar. If you guys
> (you service providers, not you Sean) are so worried about being asked
> questions in public (and short of com-priv, I see no more logical place
> for this than NANOG if "local" questions don't work], why don't you
> work on schemes to pro-actively publish performance data?  Otherwise
> you *will* get called to the table based on ping and traceroute data,
> there just ain't no other way. Either *you* do it, or others will do it
> for you, with the tools they find most appropriate, and all your
> whining will eventually be discounted. Be pro-active, or give them
> better tools. You *asked* to be in the kitchen, *you* better deliver a
> working network, including at the Internet-systemic level and not only
> in your own service provider centric environment. I suppose that if
> you guys can't figure it out yourself, there is always the opportunity
> to get it fixed by someone else. Like, ahem, regulation.

More information about the NANOG mailing list