Re: ns limitations?

[Polly Huang <huang@catarina.usc.edu>]

> There is much discussion of this in the ns-users archives (look for
> posts by Polly Huang).  Typically we've seen simulations limited more
> by memory than by processing time.  We've attacked this bottleneck
> with low-level optimizations (often very effective) and the use of
> abstraction techniques to avoid unnecessary detail in simulations when
> it's not needed.  See Huang's MASCOTS paper (available on the
> ``research using ns'' web page) for descriptions of some of the
> abstraction techniques currently available.  She's also done large
> detailed simulations---Polly, can you fill in the details about
> numbers of clients/servers/connections?

Sure. I've just checked in the large scale detailed simulation into the ns
distribution.  Check out tcl/ex/large-scale-web-traffic.tcl.  Included in
the tcl script, I gave a short description on the configuration and its
simulation scale (see below). 

This simulation utilizes the existing HTTP and TCP modules and (somewhat
abstract) dijkstra routing.  With appropriate web workload models, this
simulation creates traffic that roughly matches the 2nd order property of
our measurements taken from an ISP environment (see 'Dynamics of IP
traffic', to appear in this year's SIGCOMM)

Hope this helps.
-Polly

# Maintainer: Polly Huang <huang@isi.edu>
# Version Date: $Date: 1999/04/20 22:34:28 $
#
# @(#) $Header:
/usr/src/mash/repository/vint/ns-2/tcl/ex/large-scale-web-traffi
c.tcl,v 1.1 1999/04/20 22:34:28 polly Exp $ (USC/ISI)
#
#
# An example script that simulates large-scale web traffic. 
# See web-traffic.tcl for a smaller scale web traffic simulation.
# Some attributes:
# 1. Topology: ~460 nodes, 420 web clients, 40 web servers
# 2. Traffic: approximately 200,000 TCP connections, heavy-tailed
#             connection sizes, throughout 4200 second simulation time
# 3. Simulation scale: ~800 MB memory, ~1.5-2 hrs running on FreeBSD 3.0
#              Pentium II Xeon 450 MHz PC with 1GB physical memory
#