Actually, the traceroutes look good and are actually symmetric.
The earlier traceroute you provided was actually from a time when
the path was broken, since the first hop was out the 198.11.23
interface instead of the correct 198.11.22 interface.
David and I are continuing to investigate the performance issue.
-Bill
> Thanks to David Guevara for restoring the
> route from jsimpson-h.gsfc.nasa.gov via
> NREN!
>
> I did want to verify an earlier assertion that
> the data rates between jsimpson-h.gsfc.nasa.gov
> and hydra-atm1.colorado.edu via NREN and vBNS were
> ~80 Mbps in 6/98. The enclosed message documents
> this. See below for table of 63, 79, 78, 34, 56 Mbps for
> five tests.
>
> I know we would like to do something different with
> the network, but this research group would like to do a 512
> processor run before August 1999. This would really stress
> our code and the capabilities of the T3E, the network and
> the local data/visualization system here.
>
> The current data rates (transfers of 283 MB files) are:
>
> vBNS / NISN vBNS / NREN
> Mbps Mbps
> 02/16/99 02/23/99
> Minimum : 5.4 17.1
> Average: 8.8 20.2
> Maximum: 15.4 26.0
>
>
> These are current traceroutes in both directions:
>
> Tue Feb 23 18:12:14 MST 1999
> jsimpson> /etc/traceroute hydra-atm1.colorado.edu
> traceroute to hydra-atm1.colorado.edu (198.11.22.12), 30 hops max, 96 byte packets
> 1 169.154.161.1 (169.154.161.1) 4.247 ms 4.238 ms 4.2 ms (4.0/4.000)
> 2 rtr-atm-a.gsfc.nasa.gov (169.154.1.1) 4.686 ms 4.588 ms 4.797 ms (4.0/4.000)
> 3 198.12.123.41 (198.12.123.41) 5.632 ms 5.587 ms 4.722 ms (5.0/4.667)
> 4 chi-gsfc.nren.nasa.gov (198.10.64.50) 23.617 ms 23.872 ms 23.899 ms (23.0/23.000)
> 5 204.147.128.197 (204.147.128.197) 24.882 ms 24.873 ms 25.037 ms (24.0/24.333)
> 6 cs-atm0-0-12.dnj.vbns.net (204.147.129.186) 49.871 ms 49.103 ms 49.648 ms (49.0/49.000)
> 7 204.131.62.37 (204.131.62.37) 50.696 ms 50.71 ms 50.712 ms (50.0/50.000)
> 8 hydra-atm1.Colorado.EDU (198.11.22.12) 51.325 ms 51.637 ms 51.177 ms (51.0/51.000)
> jsimpson>
>
> Tue Feb 23 18:12:14 MST 1999
> hydra 236# traceroute jsimpson-h.gsfc.nasa.gov
> traceroute to jsimpson-h.gsfc.nasa.gov (169.154.161.43), 30 hops max, 40 byte packets
> 1 198.11.22.25 (198.11.22.25) 30 ms 30 ms 30 ms
> 2 204.131.62.38 (204.131.62.38) 30 ms 30 ms 30 ms
> 3 cs-atm0-0-14.dng.vbns.net (204.147.129.185) 30 ms 30 ms 30 ms
> 4 204.147.128.198 (204.147.128.198) 30 ms 29 ms 30 ms
> 5 gsfc-chi.nren.nasa.gov (198.10.64.49) 51 ms 47 ms 47 ms
> 6 198.12.123.42 (198.12.123.42) 47 ms 48 ms 47 ms
> 7 rtr-900-giga-a.gsfc.nasa.gov (169.154.32.3) 48 ms 47 ms 47 ms
> 8 jsimpson-h.gsfc.nasa.gov (169.154.161.43) 50 ms 49 ms 49 ms
>
> These is a traceroute from 06/98: Does anyone have one in the opposite direction?
> Is the fact that we have an asymmetrical route degrading the performance? Or is
> it that we have switched from rtr-hpcc-ea.nasa.atd.net to rtr-900-giga-a.gsfc.nasa.gov,
> (or seem to be taking a different path through NASA)?
>
> Thu Jun 11 16:10:20 1998
> traceroute to jsimpson-h.gsfc.nasa.gov (169.154.161.43), 30 hops max, 40 byte packets
> 1 cuatm-gw-arp.Colorado.EDU (198.11.23.25) 1 ms 1 ms 1 ms
> 2 204.131.62.38 (204.131.62.38) 3 ms 3 ms 3 ms
> 3 cs-atm0-0-14.dng.vbns.net (204.147.129.185) 27 ms (ttl=254!) 27 ms (ttl=254!) 26 ms (ttl=254!)
> 4 204.147.128.198 (204.147.128.198) 32 ms (ttl=253!) 28 ms (ttl=253!) 28 ms (ttl=253!)
> 5 198.10.64.129 (198.10.64.129) 43 ms (ttl=252!) 42 ms (ttl=252!) 43 ms (ttl=252!)
> 6 rtr-hpcc-ea.nasa.atd.net (198.119.5.251) 49 ms (ttl=249!) 46 ms (ttl=249!) 46 ms (ttl=249!)
> 7 * * *
> 8 jsimpson-h.gsfc.nasa.gov (169.154.161.43) 149 ms (ttl=249!) 48 ms (ttl=249!) 49 ms (ttl=249!)
>
>
> Thank you for the engineering, and all the assistance.
>
> Anne Hammond
>
>
>
>
>
> ----- Begin Included Message -----
>
> >From a-malagoli@uchicago.edu Thu Jun 11 15:44:55 1998
> Reply-To: <a-malagoli@uchicago.edu>
> From: "Andrea Malagoli" <a-malagoli@uchicago.edu>
> To: "Anne Hammond" <hammond@solarz.Colorado.EDU>
> Subject: RE: jsimpson file transfer results
> Date: Thu, 11 Jun 1998 16:40:12 -0700
> MIME-Version: 1.0
> Content-Type: text/plain; charset="iso-8859-1"
> Content-Transfer-Encoding: 7bit
> X-Priority: 3 (Normal)
> X-MSMail-Priority: Normal
> X-Mailer: Microsoft Outlook 8.5, Build 4.71.2173.0
> Importance: Normal
> X-MimeOLE: Produced By Microsoft MimeOLE V4.72.2106.4
> Content-Length: 3315
>
>
> Anne, are we at Chicago connected as well now ?
> Should we be testing ?
>
> - Andrea
>
> > -----Original Message-----
> > From: Anne Hammond [mailto:hammond@solarz.Colorado.EDU]
> > Sent: Thursday, June 11, 1998 1:04 PM
> > To: a-malagoli@uchicago.edu; mambretti@uchicago.edu
> > Cc: hammond@solarz.Colorado.EDU
> > Subject: jsimpson file transfer results
> >
> >
> > On ttcp tests, we saw 80 Mbps, so I think this is
> > the max transfer rate we will see (results below).
> >
> > Anyway, I am pleased with this.
> >
> > Randy Hudson has transferred the CAVE libraries to CU,
> > so we need to plan what we would like to do with a
> > CAVE to CAVE application.
> >
> > Anne
> >
> >
> > ----- Begin Included Message -----
> >
> > >From hammond Thu Jun 11 13:44:06 1998
> > Date: Thu, 11 Jun 1998 13:43:27 -0600 (MDT)
> > From: Anne Hammond <hammond@solarz.Colorado.EDU>
> > To: dguevara@ciocc.arc.nasa.gov
> > Subject: jsimpson file transfer results:
> > Cc: freeman@ciocc.arc.nasa.gov, mdeixler@mail.arc.nasa.gov,
> > shultz@nsipo.nasa.gov, brummell@solarz.Colorado.EDU,
> > hammond@solarz.Colorado.EDU, jtoomre@solarz.Colorado.EDU
> > Content-Length: 2032
> >
> > Hi David,
> >
> > Thank you for all you have been doing in setting up the routing
> > and working out the various bugs. Your emails have been archived to
> > lcd-www.colorado.edu:/nren if anyone needs to refer to earlier messages.
> >
> > Here are initial file transfer results from a sockets-based program
> > transferring the output of 256^2 x 128 from jsimpson.gsfc.nasa.gov:
> >
> > hydra.colorado.edu hydra-atm1.colorado.edu
> > Mbps Mbps
> > 6.3 63
> > 8.3 79
> > 7.7 78
> > 5.6 34
> > 6.6 56
> >
> >
> > We will soon have 512^2 x 128 files to transfer. This small test
> > shows a 10x increase. A large part of the time must be
> > in the filesystem overhead, so as we see larger files we should be
> > able to show a sustained 80 Mbps rate.
> >
> > If you want to monitor any of this, please let us know. I can
> > check on our end what the schedule for transfer might be.
> >
> > Anne Hammond
> >
> >
> >
> >
> > To hydra.colorado.edu:
> > jsimpson> !18
> > /tmp/brummell/sendem -v * hammond@hydra.colorado.edu:/vast/hammond/sendem
> > density: 33554432 bytes sent in 42.69 seconds (785981 bytes/second)
> > temperature: 33554432 bytes sent in 32.44 seconds (1034203 bytes/second)
> > u: 33554432 bytes sent in 35.01 seconds (958450 bytes/second)
> > v: 33554432 bytes sent in 48.22 seconds (695799 bytes/second)
> > w: 33554432 bytes sent in 40.41 seconds (830326 bytes/second)
> > jsimpson>
> >
> > To hydra-atm1.colorado.edu:
> > jsimpson> !/tmp
> > /tmp/brummell/sendem -v *
> > hammond@hydra-atm1.colorado.edu:/vast/hammond/sendem
> > density: 33554432 bytes sent in 4.25 seconds (7891962 bytes/second)
> > temperature: 33554432 bytes sent in 3.41 seconds (9853217 bytes/second)
> > u: 33554432 bytes sent in 3.43 seconds (9795015 bytes/second)
> > v: 33554432 bytes sent in 7.78 seconds (4312945 bytes/second)
> > w: 33554432 bytes sent in 4.79 seconds (6999289 bytes/second)
> > jsimpson> bc
> >
> >
> >
> > ----- End Included Message -----
> >
> >
> >
> > ----- End Included Message -----
> >
>
>
> ----- End Included Message -----