wiki:PlasticSlices/Continuation/Round9

Version 3 (modified by Josh Smift, 8 years ago) (diff)

--

Continuation Round 9

In this round, the plan was to run all ten slices on all seven sites, for a week.

Wisconsin suffered a major outage in the middle of the week, affecting all of the slices that they're a part of.

The raw logs are at http://www.gpolab.bbn.com/plastic-slices/continuation/round-9/.

plastic-101

SteadyPing, using interval=.005, and this table of client/server pairs:

client server server address
ganel.gpolab.bbn.com planetlab5.clemson.edu server=10.42.101.105
planetlab4.clemson.edu plnode2.cip.gatech.edu server=10.42.101.101
plnode1.cip.gatech.edu pl5.myplc.grnoc.iu.edu server=10.42.101.73
pl4.myplc.grnoc.iu.edu of-planet2.stanford.edu server=10.42.101.91
of-planet1.stanford.edu pl02.cs.washington.edu server=10.42.101.81
pl01.cs.washington.edu wings-openflow-3.wail.wisc.edu server=10.42.101.96
wings-openflow-2.wail.wisc.edu gardil.gpolab.bbn.com server=10.42.101.52

Commands run on each client

echo "server=$server"
sudo ping -q -i .005 -s $((1500-8-20)) $server

Results

Generated from the logs with

subnet=101
for host in $(awk 'NR%2==1' ~/plastic-slices/logins/logins-plastic-$subnet.txt | sed -r -e 's/.+@//') ; do echo -e "$host:\n\n{{{" ; grep -A 2 statistics pgenigpolabbbncom_plastic$subnet\@$host.log ; echo -e "}}}\n" ; done

ganel.gpolab.bbn.com:

--- 10.42.101.105 ping statistics ---
60616348 packets transmitted, 60101723 received, +4555 errors, 0% packet loss, time 603873971ms
rtt min/avg/max/mdev = 137.979/138.803/29421.397/74.601 ms, pipe 3868

planetlab4.clemson.edu:

--- 10.42.101.101 ping statistics ---
60428426 packets transmitted, 54618968 received, +4576 errors, 9% packet loss, time 604322150ms
rtt min/avg/max/mdev = 5.320/56.341/24832.987/38.795 ms, pipe 2486

plnode1.cip.gatech.edu:

--- 10.42.101.73 ping statistics ---
63749841 packets transmitted, 63177150 received, +4782 errors, 0% packet loss, time 603992529ms
rtt min/avg/max/mdev = 37.118/38.150/42375.436/75.287 ms, pipe 4211

pl4.myplc.grnoc.iu.edu:

--- 10.42.101.91 ping statistics ---
60563865 packets transmitted, 59943291 received, +5993 errors, 1% packet loss, time 603979365ms
rtt min/avg/max/mdev = 102.673/102.975/14960.036/33.122 ms, pipe 1414

of-planet1.stanford.edu:

--- 10.42.101.81 ping statistics ---
59627849 packets transmitted, 58755149 received, +6237 errors, 1% packet loss, time 603972071ms
rtt min/avg/max/mdev = 153.075/154.404/13256.334/43.909 ms, pipe 1356

pl01.cs.washington.edu:

--- 10.42.101.96 ping statistics ---
57466511 packets transmitted, 32971249 received, +257541 errors, 42% packet loss, time 603931891ms
rtt min/avg/max/mdev = 59.706/61.876/10494.960/37.911 ms, pipe 1009

wings-openflow-2.wail.wisc.edu:

Comments

We didn't get any log data from wings-openflow-2.wail.wisc.edu, and high packet loss from pl01.cs.washington.edu to 1042.101.96 (on wings-openflow-3.wail.wisc.edu), due to the outage at Wisconsin.

The 9% packet loss from planetlab4.clemson.edu to 10.42.101.101 (on plnode2.cip.gatech.edu) is unexpectedly high; we haven't investigated the cause.

plastic-102

SteadyPing, using interval=.005, and this table of client/server pairs:

client server server address
navis.gpolab.bbn.com planetlab4.clemson.edu server=10.42.102.104
planetlab5.clemson.edu plnode1.cip.gatech.edu server=10.42.102.100
plnode2.cip.gatech.edu pl4.myplc.grnoc.iu.edu server=10.42.102.72
pl5.myplc.grnoc.iu.edu of-planet2.stanford.edu server=10.42.102.91
of-planet1.stanford.edu pl01.cs.washington.edu server=10.42.102.80
pl02.cs.washington.edu wings-openflow-2.wail.wisc.edu server=10.42.102.95
wings-openflow-3.wail.wisc.edu bain.gpolab.bbn.com server=10.42.102.54

Commands run on each client

echo "server=$server"
sudo ping -q -i .005 -s $((1500-8-20)) $server

Results

Generated from the logs with

subnet=102
for host in $(awk 'NR%2==1' ~/plastic-slices/logins/logins-plastic-$subnet.txt | sed -r -e 's/.+@//') ; do echo -e "$host:\n\n{{{" ; grep -A 2 statistics pgenigpolabbbncom_plastic$subnet\@$host.log ; echo -e "}}}\n" ; done

navis.gpolab.bbn.com:

--- 10.42.102.104 ping statistics ---
60388197 packets transmitted, 59890888 received, +4716 errors, 0% packet loss, time 603564330ms
rtt min/avg/max/mdev = 165.433/165.715/19394.280/32.904 ms, pipe 1793

planetlab5.clemson.edu:

--- 10.42.102.100 ping statistics ---
99111985 packets transmitted, 98010857 received, +4419 errors, 1% packet loss, time 603621429ms
rtt min/avg/max/mdev = 4.902/6.045/4318.493/7.668 ms, pipe 687

plnode2.cip.gatech.edu:

--- 10.42.102.72 ping statistics ---
36404450 packets transmitted, 35859984 received, +4773 errors, 1% packet loss, time 603602380ms
rtt min/avg/max/mdev = 143.039/175.294/17996.136/56.039 ms, pipe 1080

pl5.myplc.grnoc.iu.edu:

--- 10.42.102.91 ping statistics ---
62114299 packets transmitted, 61520938 received, +5814 errors, 0% packet loss, time 603590568ms
rtt min/avg/max/mdev = 77.454/77.687/16509.553/22.526 ms, pipe 1475

of-planet1.stanford.edu:

--- 10.42.102.80 ping statistics ---
57598200 packets transmitted, 56930142 received, +5607 errors, 1% packet loss, time 603580960ms
rtt min/avg/max/mdev = 20.442/21.395/13855.638/14.768 ms, pipe 1274

pl02.cs.washington.edu:

--- 10.42.102.95 ping statistics ---
56998165 packets transmitted, 31861651 received, +257001 errors, 44% packet loss, time 603573426ms
rtt min/avg/max/mdev = 59.711/61.799/17626.355/39.431 ms, pipe 1535

wings-openflow-3.wail.wisc.edu:

Comments

We didn't get any log data from wings-openflow-3.wail.wisc.edu, and high packet loss from pl02.cs.washington.edu to 1042.101.95 (on wings-openflow-2.wail.wisc.edu), due to the outage at Wisconsin.

The absence of packet loss between Clemson and GT here, compared to plastic-101, is somewhat unexpected; we haven't investigated it.

plastic-103

SteadyPerf TCP, using port=5103, time=518400, and this table of client/server pairs:

client server server address
sardis.gpolab.bbn.com of-planet1.stanford.edu server=10.42.103.90
of-planet2.stanford.edu pl01.cs.washington.edu server=10.42.103.80
pl02.cs.washington.edu ganel.gpolab.bbn.com server=10.42.103.51

One-time prep commands run on each client and server

sudo yum -y install iperf

Commands run on each server

echo "server=$server"
nice -n 19 iperf -B $server -p 5103 -s -i 1

Commands run on each client

echo "server=$server"
nice -n 19 iperf -c $server -p 5103 -t 518400

Results

Generated with

subnet=103
for host in $(awk 'NR%2==1' ~/plastic-slices/logins/logins-plastic-$subnet.txt | sed -r -e 's/.+@//') ; do echo -e "$host:\n\n{{{" ; grep -a -A 5 -B 1 "Client connecting" pgenigpolabbbncom_plastic$subnet\@$host.log ; echo -e "}}}\n" ; done

and then edited slightly to remove artifacts (like control characters, my prompt, etc).

sardis.gpolab.bbn.com:

------------------------------------------------------------
Client connecting to 10.42.103.90, TCP port 5103
TCP window size: 16.0 KByte (default)
------------------------------------------------------------
[  3] local 10.42.103.53 port 60939 connected with 10.42.103.90 port 5103
write2 failed: No route to host
[  3]  0.0-359374.0 sec    104 GBytes  2.48 Mbits/sec
--
------------------------------------------------------------
Client connecting to 10.42.103.90, TCP port 5103
TCP window size: 16.0 KByte (default)
------------------------------------------------------------
[  3] local 10.42.103.53 port 48795 connected with 10.42.103.90 port 5103
write2 failed: No route to host
[  3]  0.0-62326.2 sec  17.6 GBytes  2.43 Mbits/sec

of-planet2.stanford.edu:

------------------------------------------------------------
Client connecting to 10.42.103.80, TCP port 5103
TCP window size: 16.0 KByte (default)
------------------------------------------------------------
[  3] local 10.42.103.91 port 36826 connected with 10.42.103.80 port 5103
[ ID] Interval       Transfer     Bandwidth
[  3]  0.0-423883.1 sec  17561894255500267520 bits  0.00 (null)s/sec
--
------------------------------------------------------------
Client connecting to 10.42.103.80, TCP port 5103
TCP window size: 16.0 KByte (default)
------------------------------------------------------------
[  3] local 10.42.103.91 port 34411 connected with 10.42.103.80 port 5103
[ ID] Interval       Transfer     Bandwidth
[  3]  0.0-179565.0 sec  13485676214369468416 bits  0.00 (null)s/sec

pl02.cs.washington.edu:

------------------------------------------------------------
Client connecting to 10.42.103.51, TCP port 5103
TCP window size: 16.0 KByte (default)
------------------------------------------------------------
[  3] local 10.42.103.81 port 46683 connected with 10.42.103.51 port 5103
write2 failed: No route to host
[  3]  0.0-486223.7 sec    127 GBytes  2.23 Mbits/sec

Comments

We've seen the "0.00 (null)s/sec" business from iperf TCP summaries before; traffic is clearly flowing, so we're not sure what causes that.

None of these ran to completion, due to temporary network failures (the "no route to host" errors). It'd be nice to have a more resilient TCP traffic generator.

plastic-104

SteadyPerf UDP, using port=5104, time=518400, rate=11, and this table of client/server pairs:

client server server address
ganel.gpolab.bbn.com wings-openflow-2.wail.wisc.edu server=10.42.104.95
pl5.myplc.grnoc.iu.edu sardis.gpolab.bbn.com server=10.42.104.53

One-time prep commands run on each client and server

sudo yum -y install iperf

Commands run on each server

echo "server=$server"
nice -n 19 iperf -u -B $server -p 5104 -s -i 1

Commands run on each client

echo "server=$server"
nice -n 19 iperf -u -c $server -p 5104 -t 518400 -b 11M

Results

Generated with

subnet=104
for host in $(awk 'NR%2==1' ~/plastic-slices/logins/logins-plastic-$subnet.txt | sed -r -e 's/.+@//') ; do echo -e "$host:\n\n{{{" ; grep -A 10 -B 1 "Client connecting" pgenigpolabbbncom_plastic$subnet\@$host.log ; echo -e "}}}\n" ; done

and then edited slightly to remove artifacts (like control characters, my prompt, etc).

ganel.gpolab.bbn.com:

------------------------------------------------------------
Client connecting to 10.42.104.95, UDP port 5104
Sending 1470 byte datagrams
UDP buffer size:   110 KByte (default)
------------------------------------------------------------
[  3] local 10.42.104.51 port 41464 connected with 10.42.104.95 port 5104
[  3]  0.0-518400.0 sec    664 GBytes  11.0 Mbits/sec
[  3] Sent 484931597 datagrams
[  3] WARNING: did not receive ack of last datagram after 10 tries.

pl5.myplc.grnoc.iu.edu:

------------------------------------------------------------
Client connecting to 10.42.104.53, UDP port 5104
Sending 1470 byte datagrams
UDP buffer size:   110 KByte (default)
------------------------------------------------------------
[  3] local 10.42.104.73 port 54328 connected with 10.42.104.53 port 5104
[  3]  0.0-518400.0 sec    664 GBytes  11.0 Mbits/sec
[  3] Sent 484932280 datagrams
[  3] Server Report:
[  3]  0.0-518390.2 sec    657 GBytes  10.9 Mbits/sec  0.017 ms 5368914/484932279 (1.1%)
[  3]  0.0-518390.2 sec  16126 datagrams received out-of-order

Comments

We didn't get back a server report from ganel.gpolab.bbn.com, due to the outage at Wisconsin. (It was connecting to 10.42.104.95.)

We haven't analyzed whether the packet loss from Indiana to BBN was consistent throughout the run, or due to a single outage, or what.

plastic-105

SteadyPerf UDP, using port=5105, time=518400, rate=6, and this table of client/server pairs:

client server server address
wings-openflow-2.wail.wisc.edu planetlab5.clemson.edu server=10.42.105.105
planetlab4.clemson.edu navis.gpolab.bbn.com server=10.42.105.55
bain.gpolab.bbn.com plnode2.cip.gatech.edu server=10.42.105.101
plnode1.cip.gatech.edu wings-openflow-3.wail.wisc.edu server=10.42.105.96

One-time prep commands run on each client and server

sudo yum -y install iperf

Commands run on each server

echo "server=$server"
nice -n 19 iperf -u -B $server -p 5105 -s -i 1

Commands run on each client

echo "server=$server"
nice -n 19 iperf -u -c $server -p 5105 -t 518400 -b 6M

Results

Generated with

subnet=105
for host in $(awk 'NR%2==1' ~/plastic-slices/logins/logins-plastic-$subnet.txt | sed -r -e 's/.+@//') ; do echo -e "$host:\n\n{{{" ; grep -A 10 -B 1 "Client connecting" pgenigpolabbbncom_plastic$subnet\@$host.log ; echo -e "}}}\n" ; done

and then edited slightly to remove artifacts (like control characters, my prompt, etc).

wings-openflow-2.wail.wisc.edu:

------------------------------------------------------------
Client connecting to 10.42.105.105, UDP port 5105
Sending 1470 byte datagrams
UDP buffer size:  110 KByte (default)
------------------------------------------------------------
[  3] local 10.42.105.95 port 60158 connected with 10.42.105.105 port 5105

planetlab4.clemson.edu:

------------------------------------------------------------
Client connecting to 10.42.105.55, UDP port 5105
Sending 1470 byte datagrams
UDP buffer size:  110 KByte (default)
------------------------------------------------------------
[  3] local 10.42.105.104 port 41408 connected with 10.42.105.55 port 5105
[ ID] Interval       Transfer     Bandwidth
[  3]  0.0-518400.0 sec   362 GBytes  6.00 Mbits/sec
[  3] Sent 264489797 datagrams
[  3] Server Report:
[  3]  0.0-518394.5 sec   358 GBytes  5.94 Mbits/sec   0.010 ms 2818608/264489796 (1.1%)
[  3]  0.0-518394.5 sec  8277 datagrams received out-of-order

bain.gpolab.bbn.com:

------------------------------------------------------------
Client connecting to 10.42.105.101, UDP port 5105
Sending 1470 byte datagrams
UDP buffer size:   110 KByte (default)
------------------------------------------------------------
[  3] local 10.42.105.54 port 37129 connected with 10.42.105.101 port 5105
[  3]  0.0-518400.0 sec    362 GBytes  6.00 Mbits/sec
[  3] Sent 264487534 datagrams
[  3] Server Report:
[  3]  0.0-518396.4 sec    357 GBytes  5.92 Mbits/sec  3.197 ms 3455085/264487534 (1.3%)
[  3]  0.0-518396.4 sec  7492 datagrams received out-of-order

plnode1.cip.gatech.edu:

------------------------------------------------------------
Client connecting to 10.42.105.96, UDP port 5105
Sending 1470 byte datagrams
UDP buffer size:   122 KByte (default)
------------------------------------------------------------
[  3] local 10.42.105.100 port 47829 connected with 10.42.105.96 port 5105
[  3]  0.0-518400.0 sec    362 GBytes  6.00 Mbits/sec
[  3] Sent 264473352 datagrams
[  3] WARNING: did not receive ack of last datagram after 10 tries.

Comments

We didn't get any log data from wings-openflow-2.wail.wisc.edu, due to the outage at Wisconsin.

We didn't get back a server report from plnode1.cip.gatech.edu, due to the outage at Wisconsin. (It was connecting to 10.42.104.96.)

We haven't analyzed whether the packet loss in the others was consistent throughout the run, or due to a single outage, or what.

plastic-106

SteadyPerf TCP, using port=5106, time=518400, and this table of client/server pairs:

client server server address
planetlab5.clemson.edu wings-openflow-2.wail.wisc.edu server=10.42.106.95
wings-openflow-3.wail.wisc.edu plnode1.cip.gatech.edu server=10.42.106.100
plnode2.cip.gatech.edu bain.gpolab.bbn.com server=10.42.106.54
navis.gpolab.bbn.com planetlab4.clemson.edu server=10.42.106.104

One-time prep commands run on each client and server

sudo yum -y install iperf

Commands run on each server

echo "server=$server"
nice -n 19 iperf -B $server -p 5106 -s -i 1

Commands run on each client

echo "server=$server"
nice -n 19 iperf -c $server -p 5106 -t 518400

Results

Generated with

subnet=106
for host in $(awk 'NR%2==1' ~/plastic-slices/logins/logins-plastic-$subnet.txt | sed -r -e 's/.+@//') ; do echo -e "$host:\n\n{{{" ; grep -A 5 -B 1 "Client connecting" pgenigpolabbbncom_plastic$subnet\@$host.log ; echo -e "}}}\n" ; done

and then edited slightly to remove artifacts (like control characters, my prompt, etc).

planetlab5.clemson.edu:

------------------------------------------------------------
Client connecting to 10.42.106.95, TCP port 5106
TCP window size: 16.0 KByte (default)
------------------------------------------------------------
[  3] local 10.42.106.105 port 55591 connected with 10.42.106.95 port 5106
[ ID] Interval       Transfer     Bandwidth
[  3]  0.0-518400.0 sec  14017702187097243648 bits  0.00 (null)s/sec

wings-openflow-3.wail.wisc.edu:

------------------------------------------------------------
Client connecting to 10.42.106.100, TCP port 5106
TCP window size: 16.0 KByte (default)
------------------------------------------------------------
[  3] local 10.42.106.96 port 32950 connected with 10.42.106.100 port 5106

plnode2.cip.gatech.edu:

------------------------------------------------------------
Client connecting to 10.42.106.54, TCP port 5106
TCP window size: 16.0 KByte (default)
------------------------------------------------------------
[  3] local 10.42.106.101 port 46041 connected with 10.42.106.54 port 5106
write2 failed: No route to host
[  3]  0.0-486087.3 sec  16.7 GBytes    296 Kbits/sec

navis.gpolab.bbn.com:

------------------------------------------------------------
Client connecting to 10.42.106.104, TCP port 5106
TCP window size: 16.0 KByte (default)
------------------------------------------------------------
[  3] local 10.42.106.55 port 60960 connected with 10.42.106.104 port 5106
write2 failed: No route to host
[  3]  0.0-486070.0 sec    171 GBytes  3.02 Mbits/sec

Comments

We've seen the "0.00 (null)s/sec" business from iperf TCP summaries before; traffic is clearly flowing, so we're not sure what causes that.

We didn't get any log data from wings-openflow-3.wail.wisc.edu, due to the outage at Wisconsin.

Neither of the others ran to completion, due to temporary network failures (the "no route to host" errors). It'd be nice to have a more resilient TCP traffic generator.

plastic-107

SteadyPerf TCP, using port=5107, time=518400, and this table of client/server pairs:

client server server address
planetlab5.clemson.edu pl4.myplc.grnoc.iu.edu server=10.42.107.72
pl5.myplc.grnoc.iu.edu plnode2.cip.gatech.edu server=10.42.107.101
plnode1.cip.gatech.edu pl02.cs.washington.edu server=10.42.107.81
pl01.cs.washington.edu planetlab4.clemson.edu server=10.42.107.104

One-time prep commands run on each client and server

sudo yum -y install iperf

Commands run on each server

echo "server=$server"
nice -n 19 iperf -B $server -p 5107 -s -i 1

Commands run on each client

echo "server=$server"
nice -n 19 iperf -c $server -p 5107 -t 518400

Results

Generated with

subnet=107
for host in $(awk 'NR%2==1' ~/plastic-slices/logins/logins-plastic-$subnet.txt | sed -r -e 's/.+@//') ; do echo -e "$host:\n\n{{{" ; grep -A 5 -B 1 "Client connecting" pgenigpolabbbncom_plastic$subnet\@$host.log ; echo -e "}}}\n" ; done

and then edited slightly to remove artifacts (like control characters, my prompt, etc).

planetlab5.clemson.edu:

------------------------------------------------------------
Client connecting to 10.42.107.72, TCP port 5107
TCP window size: 16.0 KByte (default)
------------------------------------------------------------
[  3] local 10.42.107.105 port 50196 connected with 10.42.107.72 port 5107
[ ID] Interval       Transfer     Bandwidth
[  3]  0.0-518400.0 sec  9618427753724442624 bits  0.00 (null)s/sec

pl5.myplc.grnoc.iu.edu:

------------------------------------------------------------
Client connecting to 10.42.107.101, TCP port 5107
TCP window size: 16.0 KByte (default)
------------------------------------------------------------
[  3] local 10.42.107.73 port 38928 connected with 10.42.107.101 port 5107
write2 failed: No route to host
[  3]  0.0-486026.4 sec    143 GBytes  2.53 Mbits/sec

plnode1.cip.gatech.edu:

------------------------------------------------------------
Client connecting to 10.42.107.81, TCP port 5107
TCP window size: 16.0 KByte (default)
------------------------------------------------------------
[  3] local 10.42.107.100 port 55218 connected with 10.42.107.81 port 5107
write2 failed: No route to host
[  3]  0.0-486028.4 sec    255 GBytes  4.51 Mbits/sec

pl01.cs.washington.edu:

------------------------------------------------------------
Client connecting to 10.42.107.104, TCP port 5107
TCP window size: 16.0 KByte (default)
------------------------------------------------------------
[  3] local 10.42.107.80 port 57380 connected with 10.42.107.104 port 5107
write2 failed: No route to host
[  3]  0.0-486023.2 sec    279 GBytes  4.92 Mbits/sec

Comments

We've seen the "0.00 (null)s/sec" business from iperf TCP summaries before; traffic is clearly flowing, so we're not sure what causes that.

None of the others ran to completion, due to temporary network failures (the "no route to host" errors). It'd be nice to have a more resilient TCP traffic generator.

plastic-108

SteadyPerf UDP, using port=5108, time=518400, rate=8, and this table of client/server pairs:

client server server address
wings-openflow-3.wail.wisc.edu of-planet2.stanford.edu server=10.42.108.91
of-planet1.stanford.edu pl5.myplc.grnoc.iu.edu server=10.42.108.73
pl4.myplc.grnoc.iu.edu wings-openflow-2.wail.wisc.edu server=10.42.108.95

One-time prep commands run on each client and server

sudo yum -y install iperf

Commands run on each server

echo "server=$server"
nice -n 19 iperf -u -B $server -p 5108 -s -i 1

Commands run on each client

echo "server=$server"
nice -n 19 iperf -u -c $server -p 5108 -t 518400 -b 8M

Results

Generated with

subnet=108
for host in $(awk 'NR%2==1' ~/plastic-slices/logins/logins-plastic-$subnet.txt | sed -r -e 's/.+@//') ; do echo -e "$host:\n\n{{{" ; grep -A 10 -B 1 "Client connecting" pgenigpolabbbncom_plastic$subnet\@$host.log ; echo -e "}}}\n" ; done

and then edited slightly to remove artifacts (like control characters, my prompt, etc).

wings-openflow-3.wail.wisc.edu:

------------------------------------------------------------
Client connecting to 10.42.108.91, UDP port 5108
Sending 1470 byte datagrams
UDP buffer size:  110 KByte (default)
------------------------------------------------------------
[  3] local 10.42.108.96 port 46435 connected with 10.42.108.91 port 5108

of-planet1.stanford.edu:

------------------------------------------------------------
Client connecting to 10.42.108.73, UDP port 5108
Sending 1470 byte datagrams
UDP buffer size:  110 KByte (default)
------------------------------------------------------------
[  3] local 10.42.108.90 port 60229 connected with 10.42.108.73 port 5108
[ ID] Interval       Transfer     Bandwidth
[  3]  0.0-518400.0 sec   483 GBytes  8.00 Mbits/sec
[  3] Sent 352626691 datagrams
[  3] Server Report:
[  3]  0.0-518393.7 sec   476 GBytes  7.88 Mbits/sec   0.027 ms 5105757/352626690 (1.4%)
[  3]  0.0-518393.7 sec  16667 datagrams received out-of-order

pl4.myplc.grnoc.iu.edu:

------------------------------------------------------------
Client connecting to 10.42.108.95, UDP port 5108
Sending 1470 byte datagrams
UDP buffer size:   110 KByte (default)
------------------------------------------------------------
[  3] local 10.42.108.72 port 46159 connected with 10.42.108.95 port 5108
[  3]  0.0-518400.0 sec    483 GBytes  8.00 Mbits/sec
[  3] Sent 352649481 datagrams
[  3] WARNING: did not receive ack of last datagram after 10 tries.

Comments

We didn't get any log data from wings-openflow-3.wail.wisc.edu, due to the outage at Wisconsin.

We didn't get back a server report from pl4.myplc.grnoc.iu.edu, due to the outage at Wisconsin. (It was connecting to 10.42.108.95.)

We haven't analyzed whether the packet loss in the other was consistent throughout the run, or due to a single outage, or what.

plastic-109

SteadyPerf UDP, using port=5109, time=518400, rate=4, and this table of client/server pairs:

client server server address
gardil.gpolab.bbn.com pl5.myplc.grnoc.iu.edu server=10.42.109.73
pl4.myplc.grnoc.iu.edu pl02.cs.washington.edu server=10.42.109.81
pl01.cs.washington.edu planetlab5.clemson.edu server=10.42.109.105
planetlab4.clemson.edu of-planet1.stanford.edu server=10.42.109.90
of-planet2.stanford.edu wings-openflow-3.wail.wisc.edu server=10.42.109.96
wings-openflow-2.wail.wisc.edu ganel.gpolab.bbn.com server=10.42.109.51

One-time prep commands run on each client and server

sudo yum -y install iperf

Commands run on each server

echo "server=$server"
nice -n 19 iperf -u -B $server -p 5109 -s -i 1

Commands run on each client

echo "server=$server"
nice -n 19 iperf -u -c $server -p 5109 -t 518400 -b 4M

Results

Generated with

subnet=109
for host in $(awk 'NR%2==1' ~/plastic-slices/logins/logins-plastic-$subnet.txt | sed -r -e 's/.+@//') ; do echo -e "$host:\n\n{{{" ; grep -A 10 -B 1 "Client connecting" pgenigpolabbbncom_plastic$subnet\@$host.log ; echo -e "}}}\n" ; done

and then edited slightly to remove artifacts (like control characters, my prompt, etc).

gardil.gpolab.bbn.com:

------------------------------------------------------------
Client connecting to 10.42.109.73, UDP port 5109
Sending 1470 byte datagrams
UDP buffer size:   110 KByte (default)
------------------------------------------------------------
[  3] local 10.42.109.52 port 34116 connected with 10.42.109.73 port 5109
[  3]  0.0-518400.0 sec    241 GBytes  4.00 Mbits/sec
[  3] Sent 176326532 datagrams
[  3] Server Report:
[  3]  0.0-518397.1 sec    239 GBytes  3.96 Mbits/sec  0.005 ms 1921335/176326532 (1.1%)
[  3]  0.0-518397.1 sec  8568 datagrams received out-of-order

pl4.myplc.grnoc.iu.edu:

------------------------------------------------------------
Client connecting to 10.42.109.81, UDP port 5109
Sending 1470 byte datagrams
UDP buffer size:   110 KByte (default)
------------------------------------------------------------
[  3] local 10.42.109.72 port 42414 connected with 10.42.109.81 port 5109
[  3]  0.0-518400.0 sec    241 GBytes  4.00 Mbits/sec
[  3] Sent 176326102 datagrams
[  3] Server Report:
[  3]  0.0-518397.2 sec    239 GBytes  3.96 Mbits/sec  0.891 ms 1925666/176326102 (1.1%)
[  3]  0.0-518397.2 sec  79755 datagrams received out-of-order

pl01.cs.washington.edu:

------------------------------------------------------------
Client connecting to 10.42.109.105, UDP port 5109
Sending 1470 byte datagrams
UDP buffer size:   110 KByte (default)
------------------------------------------------------------
[  3] local 10.42.109.80 port 57597 connected with 10.42.109.105 port 5109
[  3]  0.0-518400.0 sec    241 GBytes  4.00 Mbits/sec
[  3] Sent 176247368 datagrams
[  3] Server Report:
[  3]  0.0-518395.7 sec    238 GBytes  3.95 Mbits/sec  0.100 ms 2084033/176247368 (1.2%)
[  3]  0.0-518395.7 sec  24113 datagrams received out-of-order

planetlab4.clemson.edu:

------------------------------------------------------------
Client connecting to 10.42.109.90, UDP port 5109
Sending 1470 byte datagrams
UDP buffer size:  110 KByte (default)
------------------------------------------------------------
[  3] local 10.42.109.104 port 38981 connected with 10.42.109.90 port 5109
[ ID] Interval       Transfer     Bandwidth
[  3]  0.0-518400.0 sec   241 GBytes  4.00 Mbits/sec
[  3] Sent 176326532 datagrams
[  3] Server Report:
[  3]  0.0-518397.4 sec   238 GBytes  3.94 Mbits/sec   0.019 ms 2472993/176326531 (1.4%)
[  3]  0.0-518397.4 sec  10171 datagrams received out-of-order

of-planet2.stanford.edu:

------------------------------------------------------------
Client connecting to 10.42.109.96, UDP port 5109
Sending 1470 byte datagrams
UDP buffer size:  110 KByte (default)
------------------------------------------------------------
[  3] local 10.42.109.91 port 43669 connected with 10.42.109.96 port 5109
[ ID] Interval       Transfer     Bandwidth
[  3]  0.0-518400.0 sec   241 GBytes  4.00 Mbits/sec
[  3] Sent 176324610 datagrams
[  3] WARNING: did not receive ack of last datagram after 10 tries.

wings-openflow-2.wail.wisc.edu:

------------------------------------------------------------
Client connecting to 10.42.109.51, UDP port 5109
Sending 1470 byte datagrams
UDP buffer size:  110 KByte (default)
------------------------------------------------------------
[  3] local 10.42.109.95 port 59778 connected with 10.42.109.51 port 5109

Comments

We didn't get back a server report from of-planet2.stanford.edu, due to the outage at Wisconsin. (It was connecting to 10.42.109.96.)

We didn't get any log data from wings-openflow-2.wail.wisc.edu, due to the outage at Wisconsin.

We haven't analyzed whether the packet loss in the others was consistent throughout the run, or due to a single outage, or what.

plastic-110

SteadyPerf TCP, using port=5110, time=518400, and this table of client/server pairs:

client server server address
bain.gpolab.bbn.com pl01.cs.washington.edu server=10.42.110.80
pl02.cs.washington.edu of-planet1.stanford.edu server=10.42.110.90
of-planet2.stanford.edu pl4.myplc.grnoc.iu.edu server=10.42.110.72
pl5.myplc.grnoc.iu.edu plnode1.cip.gatech.edu server=10.42.110.100
plnode2.cip.gatech.edu navis.gpolab.bbn.com server=10.42.110.55

One-time prep commands run on each client and server

sudo yum -y install iperf

Commands run on each server

echo "server=$server"
nice -n 19 iperf -B $server -p 5110 -s -i 1

Commands run on each client

echo "server=$server"
nice -n 19 iperf -c $server -p 5110 -t 518400

Results

Generated with

subnet=110
for host in $(awk 'NR%2==1' ~/plastic-slices/logins/logins-plastic-$subnet.txt | sed -r -e 's/.+@//') ; do echo -e "$host:\n\n{{{" ; grep -A 5 -B 1 "Client connecting" pgenigpolabbbncom_plastic$subnet\@$host.log ; echo -e "}}}\n" ; done

and then edited slightly to remove artifacts (like control characters, my prompt, etc).

bain.gpolab.bbn.com:

------------------------------------------------------------
Client connecting to 10.42.110.80, TCP port 5110
TCP window size: 16.0 KByte (default)
------------------------------------------------------------
[  3] local 10.42.110.54 port 59225 connected with 10.42.110.80 port 5110
write2 failed: No route to host
[  3]  0.0-485821.8 sec    346 GBytes  6.12 Mbits/sec

pl02.cs.washington.edu:

------------------------------------------------------------
Client connecting to 10.42.110.90, TCP port 5110
TCP window size: 16.0 KByte (default)
------------------------------------------------------------
[  3] local 10.42.110.81 port 46991 connected with 10.42.110.90 port 5110
write2 failed: No route to host
[  3]  0.0-359034.3 sec    937 GBytes  22.4 Mbits/sec
--
------------------------------------------------------------
Client connecting to 10.42.110.90, TCP port 5110
TCP window size: 16.0 KByte (default)
------------------------------------------------------------
[  3] local 10.42.110.81 port 55111 connected with 10.42.110.90 port 5110
write2 failed: No route to host
[  3]  0.0-63692.8 sec    170 GBytes  22.9 Mbits/sec

of-planet2.stanford.edu:

------------------------------------------------------------
Client connecting to 10.42.110.72, TCP port 5110
TCP window size: 16.0 KByte (default)
------------------------------------------------------------
[  3] local 10.42.110.91 port 59670 connected with 10.42.110.72 port 5110
[ ID] Interval       Transfer     Bandwidth
[  3]  0.0-422191.7 sec  17205709671801458688 bits  0.00 (null)s/sec
--
------------------------------------------------------------
Client connecting to 10.42.110.72, TCP port 5110
TCP window size: 16.0 KByte (default)
------------------------------------------------------------
[  3] local 10.42.110.91 port 57556 connected with 10.42.110.72 port 5110
[ ID] Interval       Transfer     Bandwidth
[  3]  0.0-179228.7 sec  13026921957609984000 bits  0.00 (null)s/sec

pl5.myplc.grnoc.iu.edu:

------------------------------------------------------------
Client connecting to 10.42.110.100, TCP port 5110
TCP window size: 16.0 KByte (default)
------------------------------------------------------------
[  3] local 10.42.110.73 port 34143 connected with 10.42.110.100 port 5110
write2 failed: No route to host
[  3]  0.0-134175.6 sec  45.2 GBytes  2.89 Mbits/sec
--
------------------------------------------------------------
Client connecting to 10.42.110.100, TCP port 5110
TCP window size: 16.0 KByte (default)
------------------------------------------------------------
[  3] local 10.42.110.73 port 55400 connected with 10.42.110.100 port 5110
write2 failed: No route to host
[  3]  0.0-240750.2 sec  83.3 GBytes  2.97 Mbits/sec

plnode2.cip.gatech.edu:

------------------------------------------------------------
Client connecting to 10.42.110.55, TCP port 5110
TCP window size: 16.0 KByte (default)
------------------------------------------------------------
[  3] local 10.42.110.101 port 52109 connected with 10.42.110.55 port 5110
write2 failed: No route to host
[  3]  0.0-134135.9 sec  8.72 GBytes    559 Kbits/sec
--
------------------------------------------------------------
Client connecting to 10.42.110.55, TCP port 5110
TCP window size: 16.0 KByte (default)
------------------------------------------------------------
[  3] local 10.42.110.101 port 35891 connected with 10.42.110.55 port 5110
write2 failed: No route to host
[  3]  0.0-240675.2 sec  14.1 GBytes    503 Kbits/sec

Comments

We've seen the "0.00 (null)s/sec" business from iperf TCP summaries before; traffic is clearly flowing, so we're not sure what causes that.

None of the others ran to completion, due to temporary network failures (the "no route to host" errors). It'd be nice to have a more resilient TCP traffic generator.