Discussion list for thrulay, a network testing tool

[stanislav shalunov <>]

Connie,

Somewhere along the line there's a huge buffer that's filled up.
Might be honest-to-goodness network congestion (buffers of several
seconds are more common than one would hope), or it might be a
measurement artifact (an internal queue inside the measurement
machine).

What delay do you observe with ping and/or owping to the BNL site when
throughput tests are not running?

Given that the delay values measured with thrulay seem to be
increasing (Is it possible to run a longer test?  At least the default
60 seconds?), chances are the buffer is internal to the client host
(the one from which thrulay, rather than thrulayd, is run); if that's
right, then there could be some non-congestive loss on the path.

                                                        --Stas

"Logg, Connie A." 
<>
 writes:

> I am trying to use thrulay to diagnose (find) low bandwidth
> problems.  Below I run it to 3 nodes, One at caltech, one at cern,
> and one at bnl.  We are assuming the bnl node is 1 gig also.  Note
> that bnl gives very low throughput and very high rtt. It is much
> must worse than cern.
> 
> Does anyone have any ideas why this might be?  All of our tests are
> show low bandwidth to bnl.
> 
> Connie
> 
> Connie Logg - Network Analyst 
> Stanford Linear Accelerator Center 
> 2575 Sand Hill Road, Menlo Park, CA 94025
> 650-926-2879, 
> 
> "If is the middle word in life, but happiness is found along the way, not 
> at the end of the road."
> 
> 
> > ______________________________________________ 
> > From:       Logg, Connie A.  
> > Sent:       Monday, March 28, 2005 4:09 PM
> > To: Cottrell, Les
> > Subject:    3 site comparison
> > 
> > 
> > 540cal@nettest2:
> >  thrulay -t 15 socrates.cacr.caltech.edu
> > # local window = 8388608B; remote window = 8388608B
> > # block size = 8192B
> > # test duration = 15s; reporting interval = 1s
> > # begin, s  end, s  Mb/s    RTT, ms: min  avg      max
> >     0.000    1.013  127.970   60.774  347.234  544.039
> >     1.013    2.005  165.705  281.654  304.308  322.964
> >     2.005    3.010  174.784  270.473  287.119  305.195
> >     3.010    4.005  182.550  251.698  274.526  297.148
> >     4.005    5.002  189.492  241.689  264.378  280.380
> >     5.002    6.016  195.973  243.144  255.918  271.018
> >     6.016    7.000  201.608  231.258  248.784  271.616
> >     7.000    8.016  207.058  219.784  242.043  262.371
> >     8.016    9.001  378.999   86.427  143.653  248.692
> >     9.001   10.007  612.159   70.501   82.888   95.179
> >    10.007   11.003  677.555   68.272   74.853   86.818
> >    11.003   12.006  730.937   60.483   68.686   77.425
> >    12.006   13.004  769.255   54.579   65.030   81.205
> >    13.004   14.001  812.013   52.381   61.406   74.485
> >    14.001   15.003  494.601   55.863   99.081  118.561
> > #   0.000   15.003  394.431   52.381  126.856  544.039
> > 
> > 
> > 541cal@nettest2:
> >  thrulay -t 15 pcgiga.cern.ch
> > # local window = 8388608B; remote window = 8388608B
> > # block size = 8192B
> > # test duration = 15s; reporting interval = 1s
> > # begin, s  end, s  Mb/s    RTT, ms: min  avg      max
> >     0.000    1.037    1.074  505.381 1186.711 1542.244
> >     1.037    2.032   22.797 1542.269 2178.259 2530.916
> >     2.032    3.006  154.542  199.308  765.646 2707.753
> >     3.006    4.017  234.742  166.105  216.948  361.916
> >     4.017    5.006  251.841  166.160  193.848  243.294
> >     5.006    6.116  156.582  168.959  212.099  594.177
> >     6.116    7.007   65.257  593.313  885.178 1215.424
> >     7.007    8.012  141.368  266.727  388.116  724.935
> >     8.012    9.041  142.460  244.362  327.710  409.416
> >     9.041   10.011  152.071  261.289  319.297  382.199
> >    10.011   11.007  147.746  261.606  317.665  403.618
> >    11.007   12.011  146.701  252.688  326.833  397.908
> >    12.011   13.006  142.232  284.193  328.283  401.246
> >    13.006   14.004  146.668  273.632  332.103  404.795
> >    14.004   15.034  150.079  290.990  327.100  385.978
> > #   0.000   15.034  137.403  166.105  360.599 2707.753
> > 
> > 
> > 542cal@nettest2:
> >  thrulay -t 15 atlasprod3.usatlas.bnl.gov
> > # local window = 8388608B; remote window = 8388608B
> > # block size = 8192B
> > # test duration = 15s; reporting interval = 1s
> > # begin, s  end, s  Mb/s    RTT, ms: min  avg      max
> >     0.000    1.057   15.753  177.788  937.399 1230.636
> >     1.057    2.000   11.459 1230.387 1577.201 2172.282
> >     2.000    3.007    9.897 2172.265 2651.757 3176.500
> >     3.007    4.020   10.609 3176.260 3657.310 4186.513
> >     4.020    5.026   11.397 3770.564 3970.239 4260.845
> >     5.026    6.031   12.261 4113.077 4463.307 4758.109
> >     6.031    7.049    6.502 4467.244 4610.031 4906.013
> >     7.049    8.068    4.308 4906.036 5220.364 5493.553
> >     8.068    9.006    5.171 5493.571 5721.198 5927.811
> >     9.006   10.021    6.072 5927.873 6157.883 6363.394
> >    10.021   11.033    6.997 6362.849 6550.688 6723.245
> >    11.033   12.047    7.816 6723.205 6893.487 7092.054
> >    12.047   13.064    8.638 7020.672 7205.746 7390.961
> >    13.064   14.002    9.574 7315.839 7440.941 7536.118
> >    14.002   15.008    6.771 6723.759 6933.584 7388.117
> > #   0.000   15.008    8.899  177.788 4465.330 7536.118
> > 5
> > 
> > Connie Logg - Network Analyst 
> > Stanford Linear Accelerator Center 
> > 2575 Sand Hill Road, Menlo Park, CA 94025
> > 650-926-2879, 
> > 
> > "If is the middle word in life, but happiness is found along the way, not 
> > at the end of the road."
> > 

-- 
Stanislav Shalunov              http://www.internet2.edu/~shalunov/

Antibiotics comprise 2% of cow feed in U.S. feedlots, by weight.