Netperf¶

Dtaht suggested that we consider netperf instead of or in addition to iperf for the following reasons:

• netperf doesn't require the server process to be running constantly, as it can be instantiated by xinetd.
• netperf reports cpu usage on client and server which is useful in CPU-constrained environments
• netperf rus tests over both TCP and UDP
• netperf has a number of tests that can be run

Users on the internet have reported higher reported throughputs using netperf than when using iperf, for the exact same setup and devices.

Porting and testing bleeding-edge netperf¶

We downloaded the subversion HEAD (http://www.netperf.org/svn/netperf2/trunk/) and created a netperf package using this source, based on the existing openwrt netperf package.

We installed this packaged netperf (netperf-latest) on two wndr3700v2 boxes, as well as netperf 2.4.4 as packaged by Ubuntu.

Ran netperf between the following: (see bottom for device specifications)

1) LAPTOP1 -> LAPTOP2

kyriakos@zenith:~/Desktop/bismark-summercamp/netperf/trunk/src$ ./netperf -H 172.31.8.115 -c -C
MIGRATED TCP STREAM TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to 172.31.8.115 (172.31.8.115) port 0 AF_INET 
Recv   Send    Send                          Utilization       Service Demand
 Socket Socket  Message  Elapsed              Send     Recv     Send    Recv
 Size   Size    Size     Time     Throughput  local    remote   local   remote
 bytes  bytes   bytes    secs.    10^6bits/s  % S      % S      us/KB   us/KB

 87380  16384  16384    10.37         4.11   0.96     0.19     19.217  3.866  

2) router -> LAPTOP2

root@gw:/tmp# netperf -H 172.31.8.115 -c -C
MIGRATED TCP STREAM TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to 172.31.8.115 (172.31.8.115) port 0 AF_INET
Recv   Send    Send                          Utilization       Service Demand
Socket Socket  Message  Elapsed              Send     Recv     Send    Recv
Size   Size    Size     Time     Throughput  local    remote   local   remote
bytes  bytes   bytes    secs.    10^6bits/s  % S      % S      us/KB   us/KB

 87380  16384  16384    10.07         4.21   29.00    1.00     565.055  19.390 

3) router - router

root@gw:/tmp# netperf -H 172.31.8.33 -c -C
MIGRATED TCP STREAM TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to 172.31.8.33 (172.31.8.33) port 0 AF_INET
Recv   Send    Send                          Utilization       Service Demand
Socket Socket  Message  Elapsed              Send     Recv     Send    Recv
Size   Size    Size     Time     Throughput  local    remote   local   remote
bytes  bytes   bytes    secs.    10^6bits/s  % S      % S      us/KB   us/KB

 87380  16384  16384    10.07         4.23   29.49    32.41    571.139  627.530 

iperf between laptop -> laptop for comparison:

kyriakos@zenith:~$ iperf -c 172.31.8.115
------------------------------------------------------------
Client connecting to 172.31.8.115, TCP port 5001
TCP window size: 16.0 KByte (default)
------------------------------------------------------------
[  3] local 10.0.2.15 port 51176 connected with 172.31.8.115 port 5001
[ ID] Interval       Transfer     Bandwidth
[  3]  0.0-10.1 sec  5.20 MBytes  4.33 Mbits/sec

LAPTOP1:
Model Name: MacBook Pro
Model Identifier: MacBookPro6,2
Processor Name: Intel Core i5
Processor Speed: 2.53 GHz
Number Of Processors: 1
Total Number Of Cores: 2
L2 Cache (per core): 256 KB
L3 Cache: 3 MB
Memory: 4 GB
Processor Interconnect Speed: 4.8 GT/s
Boot ROM Version: MBP61.0057.B0C
SMC Version (system): 1.58f16
Serial Number (system): W8023204AGY
Hardware UUID: 519F0E32-09F0-5FB2-9BEA-A5A410ECACE3
Sudden Motion Sensor:
State: Enabled

LAPTOP2