Traffic Burstiness Survey

Naslund, Steve SNaslund at
Tue Sep 11 13:52:44 UTC 2012

Bursty is a very relative thing.  It depends on the time frame you are
considering.  For example, at any given instant of time a circuit is
either carrying data or it isn't.  The network is always either 100% in
use or 100% idle if you look at it in an instantaneous fashion.  There
is also a misconception that bursty is bad.  In very fast networks, the
systems complete their transmissions very quickly so appear bursty.  In
a lower bandwidth network the system may transmit almost continuously
because there is always data queued for transmission.  This would appear
as non-bursty continuous traffic.  What you are really looking at here
is the value and effect of traffic shaping.  You should probably repose
your question and ask about traffic shaping.  In my opinion, like most
QoS mechanisms, traffic shaping is something you do as a compromise to
get the best service in a network that does not have enough bandwidth or
has high latency.  The best solution is almost always more bandwidth.

The idea of suppressing bursts is not something that I think would be
optimal.  What you are saying is really that a system has data to send
but I am not going to allow it to go out as fast as possible.  I am
going to do some kind of traffic shaping which might be more fair to
other traffic but at the end of the day, all you can do is add delay.

Steven Naslund

-----Original Message-----
From: Monia Ghobadi [mailto:monia at] 
Sent: Saturday, September 08, 2012 7:24 PM
To: NANOG at
Subject: Traffic Burstiness Survey

Dear Nanog members,

I am a PhD student at University of Toronto and I am working on traffic
burstiness in data centers. In the following I am asking two questions
raise motivation for my research. I appreciate if anyone could answer
questions to their best knowledge. *The questions are:*

1) 'Bursty' is a word with no agreed meaning. How do you define a bursty
2) If you are involved with a data center, is your data center traffic
   -- If yes,
         -- Do you think that it will be useful to supress the
in your traffic? (For example by pacing the traffic into shorter bursts)
    -- If no:
        -- Are you already supressing the burstiness? How?
         -- Would you anticipate the traffic becoming burstier in the


Monia Ghobadi
PhD Student
University of Toronto

More information about the NANOG mailing list