Home page logo
/

nanog logo nanog mailing list archives

Traffic Analysis ( was Re: Initial Route Server Stats)
From: Tim Bass <bass () dune silkroad com>
Date: Sun, 22 Oct 1995 22:57:54 -0400 (EDT)

Last one, then i'm dropping of this thread :+

mo hits the nail on the head when he states that detailed performance
numbers in complex multihomed interconnected networks are currently
unobtainable.  In any complex problem with numerous uncertain
parameters, variables, and relative metrics all moving on the wire
at varying speeds, densities, and with a given broken path 'here
or there' at any given moment...... 

Most practical, day to day network users judge traffic on the net
the same way we judge the traffic jams on the highways....

You look out over the overpass and determine the density of the
traffic, guess about the status of traffic lights and construction
zones, wonder if there was an accident and decide if getting on
the ramp is a good idea because, once you do, there is no turning
back ( for all practical purposes ).

So, you want to know how the traffic is to Washington DC from
the suburbs.... one way is to get into your car and drive the
round trip to the city and back and then estimate the traffic.
Of course, a little Miata moves faster than an semi or Station
Wagon with 3 kids and a dog; and that police or fire vehicle
gets a higher priority... BUT in general, we know something
very simple.... "if it moves slow, it must be slow for some
reason" and we generally do not like it.

So, here we are, crusin' the network and we want to understand
why we can't get to our favorite IP address or IP address of
the moment.  We PING the thing, knowing full well that it is
not Einstein's SuperSpecial Theory of Network Metrics, but
at least we can look the the numbers and say.... geezzz...
it took a looooong time, must be a lot of traffic.  Slow is
slow even in zen-metrics.  Hey, let's do a TRACEROUTE and
guess where the congestion is...

While the PHD candidates in EE and CS programs in univs'
enjoy digging into into complex network theory, the average
day-to-day running the network, answering the phone, or just
trying to get out to some site kinda cyberjocky enjoys
their only dependable tools, PING and TRACEROUTE.

World peace, a perfect universe, harmonic diverity are all
interesting topics.  So are complex network metrics and 
and space-time theory.... time travel anyone ? :-)

Maybe we need a new Internet Traffic Report Morning Show.  
Mae-east is hosed today, but will have an new OC12 next week.
NET99 is moving over to MCI but having trouble peering with
Sprint until further notice.  Packets to the convention in
the Balkans are being filtered by CIDRites at ICP....

Stay tuned and have a nice day.  Now, up to the minute
IP security tips sponsored by cisco, your partners in
datametrics and virtual resolution protocol.  But first,
MCI and the ReachOut and Spam Your Friends and Family
special offer, after these messages.

................



-- 
+--------------------------------------------------------------------------+
| Tim Bass                           | #include<campfire.h>                | 
| Principal Network Systems Engineer |       for(beer=100;beer>1;beer++){  |
| The Silk Road Group, Ltd.          |           take_one_down();          |
|                                    |           pass_it_around();         |
| http://www.silkroad.com/           |       }                             |
|                                    |  back_to_work(); /*never reached */ | 
+--------------------------------------------------------------------------+


  By Date           By Thread  

Current thread:
  • Traffic Analysis ( was Re: Initial Route Server Stats) Tim Bass (Oct 23)
[ Nmap | Sec Tools | Mailing Lists | Site News | About/Contact | Advertising | Privacy ]