routesim files
NAME
rsfiles - Description of routesim simulator file formats
SYNOPSIS
rstopology.dat
rstopology.dat.out
rstraffic.dat
holdcdf.dat
rsmonitor.dat
rslinkmonitor.log
routesim.log
configuration file
DESCRIPTION
rstopology.dat
The rstopology file describes a arbitrary network
topology for use in a routesim simulation. Each line
of the topology file specifies a network link, and all
links are unidirectional. The format of a topology
file is as follows:
num nodes
from to capacity propdelay adminweight maxcalls
.
.
.
Items are separated by a space. The first line in the
file is an integer indicating the number of nodes in
the network. Node ids are in the range 0 to num nodes
- 1. Each subsequent line defines a link and has 6
fields. from and to specify the nodes the link con-
nects. The third, fourth, and fifth fields are float-
ing point numbers. Link capacity is specified in the
third field. Propagation delay is specified by the
fourth field. An arbitrary administrative weight is in
the fifth field. The final field is the maximum number
flows that can be supported on the link (integer).
Each field, except from and to may be set to -1.0 (or
-1 for maxcalls ) to get assigned the default value.
rstopology.dat.out
This file is dumped when the -dumptop or dump-topology
options are specified to routesim. It follows the for-
mat described for rstopology.dat above.
rstraffic.dat
The rstraffic file specifies a routesim traffic matrix.
The traffic matrix specifies the arrival rate between
each source-destination pair. It also allows different
values for different times-of-day. The arrival rate is
expressed in terms of a scale value and shape value
that serve as parameters to the flow arrival distribu-
tion. The possible distributions are Poisson and
Weibull, and is specified in a routesim configuration
file or on the command line. In the case of Poisson,
only the scale value is significant, and it represents
the mean value. For the Weibull distribution, both the
scale and shape values are significant. Together they
determine the mean of the distribution. The file for-
mat is:
from to mscale mshape ascale ashape escale eshape
.
.
.
Each line describes a traffic rate from the node from
to the node to. The mscale , ascale , and escale are
floating point values that specify the scale parameters
for morning, afternoon, and evening, respectively.
mshape , ashape , and eshape are the corresponding
shape parameters.
Note: scales are the arrival rates -- NOT interarrival
times.
holdcdf.dat
The holdcdf file specifies an inverse cumulative dis-
tribution function for the flow durations. The file
format is:
num entries
u x_first
.
.
.
1 x_last
The first line specifies the number of entries that
follows. Each subsequent line is a single ( u , x )
entry, in ascending u order. The first entry should
NOT be for u = 0. But the last entry must be for u =
1. See examples below for further explanation.
rsmonitor.dat
The rsmonitor file specifies the network links that
should be monitored during simulaton. It's format is
simply:
from to
.
.
.
Each line specifies the node ids for the link end-
points. Typically only a few links (maybe just one)
are specified since the monitoring information copious.
rslinkmonitor.log
The rslinkmonitor file is the log of events on the
links specified in rsmonitor.dat
routesim.log
This file logs the progress of the routesim simulation.
It lists the phase (warmup or actual sim), current sim-
ulation time, number of flow requests, and confidence
interval for the stopping metric.
configuration file
The configuration file contains routesim options, one
per line, in long format. It may also contain comment
lines with /* ... */ delimiters. Note that there must
be a space between the '*' character and the comment
text.
EXAMPLES
rstopology.dat
16
0 1 1.000000 0.000000 0.000000 2147483647
0 3 1.000000 0.000000 0.000000 2147483647
0 4 1.000000 0.000000 0.000000 2147483647
0 12 1.000000 0.000000 0.000000 2147483647
1 0 1.000000 0.000000 0.000000 2147483647
1 2 1.000000 0.000000 0.000000 2147483647
1 5 1.000000 0.000000 0.000000 2147483647
.
.
.
rstraffic.dat
0 1 0.150000 1.000000 0.150000 1.000000 0.150000
1.000000
0 2 0.113208 1.000000 0.113208 1.000000 0.113208
1.000000
0 3 0.150000 1.000000 0.150000 1.000000 0.150000
1.000000
0 4 0.113208 1.000000 0.113208 1.000000 0.113208
1.000000
0 5 0.113208 1.000000 0.113208 1.000000 0.113208
1.000000
.
.
. 2 0 0.113208 1.000000 0.113208 1.000000 0.113208
1.000000
2 1 0.150000 1.000000 0.150000 1.000000 0.150000
1.000000
2 3 0.150000 1.000000 0.150000 1.000000 0.150000
1.000000
2 4 0.113208 1.000000 0.113208 1.000000 0.113208
1.000000
2 5 0.113208 1.000000 0.113208 1.000000 0.113208
1.000000
2 6 0.113208 1.000000 0.113208 1.000000 0.113208
1.000000
2 7 0.113208 1.000000 0.113208 1.000000 0.113208
1.000000
.
.
.
holdcdf.dat
Consider the following CDF:
F(x) =
0.0 0 <= x < 64
0.7 64 <= x < 128
0.8 128 <= x < 256
0.9 256 <= x < 512
1.0 512 <= x
The corresponding inverse CDF is:
-1
F (x) =
64 0 < u <= 0.7
128 0.7 < u <= 0.8
256 0.8 < u <= 0.9
512 0.9 < u <= 1.0
and the corresponding holdcdf.dat file is:
4
0.7 64
0.8 128
0.9 256
1.0 512
routesim.log
A typical log file looks like:
Warming up ...
847500 (0.0022 [0.0021, 0.0024]) (overutil) Time:
725.72
Warmup complete: 848011 calls
2655000 (0.0566 [0.0563, 0.0568]) (overutil) Time:
2998.96
Simulation complete. Dumping stats ...
configuration file
Sample configuration file:
topology-type file
network-model SINGLE
arrival-spec ARR_UNIFORM
arrival-dist POISSON
arrival-scale .94118
arrival-shape 1.00
traffic-multiplier 1.0
time-of-day AFTERNOON
holding-time-spec USERDEF
holding-time-shape 2.5
holding-time-scale 1.00
mean-bandwidth 0.030
bandwidth-spread 1.99000
update-trigger 100000.0000
refresh-interval 120.000000
refresh-skew 0.02
min-update-interval 0.000000
min-requests 800000
min-warmup 500000
min-sim-time 3000.0
warmup-proportion 0.25
blocking-policy NON_BLOCKING
routing-policy ON_DEMAND
routing-algorithm WIDE_SHORT
distance-function HOPCOUNT
hopcount-threshold 1
wide-short-maxhops 10
alternate-routing NO_ALT
multi-route-policy UNIQ
uniq-route-policy RANDOM
max-sig-attempts 1
prune-policy NO_PRUNE
link-state-accuracy STALE
random-seed 1013553468
confidence-level 99
sim-tolerance .05
warmup-tolerance .075
SEE ALSO
routesim(l)
AUTHOR
Anees Shaikh (University of Michigan, AT&T Labs--Research,
1997-99)
Man(1) output converted with
man2html