[brite-users] Differences between cppgen and javagen

Gilles BERTRAND gilles.bertrand at telecom-bretagne.eu
Fri Jun 13 04:23:40 EDT 2008


Dear Brite users,

I have noticed small differences between javagen and cppgen: I would 
like to report what looks like a bug in cppgen. Note that I am very new 
to BRITE, so, maybe it is not a bug but just the consequence of my lack 
of experience with BRITE, in this case I apologize for this e-mail.

I wanted to try generating a 2-level hierarchical topology using the 
Top-Down approach. As a test, I have tried to use GLP model for the AS 
level. I used the GUI to generate the configuration file (You can find 
the contents of GUI_GEN.conf at the end of this e-mail). The generation 
worked with javagen but not with cppgen (the log is provided at the end 
of this e-mail). I am new to topology generation models: does this error 
mean that GLP model should not be used for AS-level topologies?

I have tried to investigate the source of this problem. I have noticed 
that in C++/Parser.cc in the code block from line 742 to line 779 no 
case for AS GLP=12 is defined. Maybe I am totally wrong (once again, I 
am just discovering BRITE...) but I guess this may be the source of the 
error?

--
Cordialement, with kind regards,

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Gilles BERTRAND, PhD candidate
Institut TELECOM; TELECOM Bretagne; RSM department
http://www.rennes.telecom-bretagne.eu/~gbertran/ 
<http://www.rennes.telecom-bretagne.eu/%7Egbertran/>
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~



************************* GUI_GEN.conf *********************************

#This config file was generated by the BRITE GUI.

BriteConfig

BeginModel
    Name = 5         #Top Down = 5
    edgeConn = 1         #Random=1, Smallest Nonleaf = 2, Smallest Deg = 
3, k-Degree=4
    k = -1             #Only needed if edgeConn is set to K-Degree, 
otherwise use -1
    BWInter = 1         #Constant = 1, Uniform =2, HeavyTailed = 3, 
Exponential =4
    BWInterMin = 10.0
    BWInterMax = 1024.0
    BWIntra = 1         #Constant = 1, Uniform =2, HeavyTailed = 3, 
Exponential =4
    BWIntraMin = 10.0
    BWIntraMax = 1024.0
EndModel

BeginModel
    Name = 12          #Router GLP=11, AS GLP=12
    N = 3         #Number of nodes in graph
    HS = 1000         #Size of main plane (number of squares)
    LS = 100         #Size of inner planes (number of squares)
    NodePlacement = 1     #Random = 1, HeavyTailed = 2
    m = 2             #Number of neighboring node each new node connects to.
    BWDist = 1         #Constant = 1, Uniform =2, HeavyTailed = 3, 
Exponential =4
    BWMin = 10.0
    BWMax = 1024.0
     p = 0.45         #Probability of adding links
     beta = 0.64         #linear shift
EndModel


BeginModel
     Name=11          #Router GLP=11, AS GLP=12
    N = 5         #Number of nodes in graph
    HS = 1000         #Size of main plane (number of squares)
    LS = 100         #Size of inner planes (number of squares)
    NodePlacement = 1         #Random = 1, Heavy Tailed = 2
    m = 2             #Number of neighboring node each new node connects to.
    BWDist = 1         #Constant = 1, Uniform =2, HeavyTailed = 3, 
Exponential =4
    BWMin = 10.0
    BWMax = 1024.0
     p = 0.45         #Probability of adding links
     beta = 0.64         #linear shift
EndModel

BeginOutput
    BRITE = 1      #1/0=enable/disable output in BRITE format
    OTTER = 0      #1/0=enable/disable visualization in otter
    DML = 0      #1/0=enable/disable output to SSFNet's DML format
    NS = 0     #1/0=enable/disable output to NS-2
    Javasim = 0     #1/0=enable/disable output to Javasim
EndOutput


******************************* Log when trying with cppgen ****************

Place seed used: 31436 8273 31478
Connect seed used: 40654 44405 45634
Edge conn seed used: 25477 7075 46799
Grouping seed used: 10586 4301 5652
Assigment seed used: 37335 38273 5984
Bandwidth seed used: 46705 11013 31882
Parsing Top-down hierarchical model...
Invalid model type for level 1 of hierarchical topology...


More information about the brite-users mailing list