User-level ISP performance measurement

Scott Huddle huddle at
Mon Dec 1 20:37:16 UTC 1997


Check out Inverse Network Technology.


> From owner-nanog at Mon Dec  1 15:15 EST 1997
> Date: Mon, 1 Dec 1997 14:50:29 -0500 (EST)
> From: John R Levine <johnl at>
> To: nanog at
> Subject: User-level ISP performance measurement
> MIME-Version: 1.0
> Sender: owner-nanog at
> Content-Type> : > TEXT/PLAIN> ; > charset=US-ASCII> 
> Content-Length: 1422
> I hope I'm not reopening a huge can of worms here, but ...
> Some of my relatives have started a small ISP in western Vermont.  (They
> already run the local telco.) They're wondering how to compare the
> performance that their users see to the ISPs in the adjacent town.  I told
> them that there isn't anything published that'll be of much use, since the
> stuff I've seen tries to measure backbone performance and for the most part
> does so ineptly. 
> For user level stuff, it seems to me that you could pick a representative set
> of actions (visit large and small web sites, download e-mail, scan
> newsgroups), dial in via PPP every hour for a week being sure to flush out
> caches and otherwise start from a consistent user state, do the actions, time
> them all, and make some simple stats.  This would be miserable to do
> manually, but pretty straightforward to do automatically.  Is there anything
> available, either as a product or service, that does this sort of analysis? 
> One of the other hats I wear is as a director of a software testing tools
> company, and we could certainly whip up this kind of stuff in no time using
> our existing products if there were a demand for it. 
> Regards,
> John Levine, johnl at, Primary Perpetrator of "The Internet for Dummies",
> Information Superhighwayman wanna-be,, Sewer Commissioner
> Finger for PGP key, f'print = 3A 5B D0 3F D9 A0 6A A4  2D AC 1E 9E A6 36 A3 47 

More information about the NANOG mailing list