How to monitor network bandwidth in Linux using netperf tool

We have many tools available online which can help us monitor network bandwidth usage within a network between differnt Linux machines.

Brief information on my setup:

Server

My server IP is 192.169.32.15 where I have 2 bond interfaces as below

# ifconfig bond0
bond0     Link encap:Ethernet  HWaddr 00:17:A4:77:00:38
          inet addr:192.169.32.15  Bcast:192.169.32.31  Mask:255.255.255.224
          UP BROADCAST RUNNING MASTER MULTICAST  MTU:1500  Metric:1
          RX packets:9042150 errors:0 dropped:4553 overruns:0 frame:0
          TX packets:2277574 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:0
          RX bytes:13654658052 (13022.0 Mb)  TX bytes:153728769 (146.6 Mb)
# ifconfig bond1
bond1     Link encap:Ethernet  HWaddr 38:63:BB:2F:E5:A9
          inet addr:192.169.32.37  Bcast:192.169.32.63  Mask:255.255.255.224
          UP BROADCAST RUNNING MASTER MULTICAST  MTU:1500  Metric:1
          RX packets:28334533 errors:0 dropped:2 overruns:0 frame:0
          TX packets:2066047 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:0
          RX bytes:42898423316 (40911.1 Mb)  TX bytes:136362958 (130.0 Mb)

The ethernet speed assigned to both interfaces eth0 and eth1 under bond0 is 1 GB

# ethtool eth0 | grep -i Speed
        Speed: 1000Mb/s

# ethtool eth1 | grep -i Speed
        Speed: 1000Mb/s

so if we use bond0 for our network operations then ethernet bandwidth must not exceed 1 GB

The ethernet speed assigned to both interfaces eth2 and eth3 under bond1 is 3 GB

# ethtool eth2 | grep -i Speed
        Speed: 3000Mb/s

# ethtool eth3 | grep -i Speed
        Speed: 3000Mb/s

so if we use bond1 for our network operations then ethernet bandwidth must not exceed 3 GB

 

Client

My clinet node is 192.169.32.17 where as well I have 2 bond interfaces in the same subnet as my server

# ifconfig  bond0
bond0     Link encap:Ethernet  HWaddr 00:17:A4:77:00:44
          inet addr:192.169.32.17  Bcast:192.169.32.31  Mask:255.255.255.224
          UP BROADCAST RUNNING MASTER MULTICAST  MTU:1500  Metric:1
          RX packets:2263768 errors:0 dropped:5390 overruns:0 frame:0
          TX packets:8626819 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:0
          RX bytes:150407672 (143.4 Mb)  TX bytes:12455852277 (11878.8 Mb)
# ifconfig  bond1
bond1     Link encap:Ethernet  HWaddr 38:63:BB:2F:56:C1
          inet addr:192.169.32.38  Bcast:192.169.32.63  Mask:255.255.255.224
          UP BROADCAST RUNNING MASTER MULTICAST  MTU:1500  Metric:1
          RX packets:2067482 errors:0 dropped:16 overruns:0 frame:0
          TX packets:28335782 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:0
          RX bytes:136453804 (130.1 Mb)  TX bytes:41070143636 (39167.5 Mb)

The ethernet speed assigned to bond0 and bond1 is same as server i.e. 1 GB and 3 GB respectively

Let us start our test

Installed netperf on both the machines

NOTE: I would recommend to disable the firewall on both server and client for the monitoring purpose as at times your firewall might stop netperf from sending traffic to client machine

On Server

server # rpm -Uvh /tmp/netperf-2.6.0-3.3.x86_64.rpm
warning: /tmp/netperf-2.6.0-3.3.x86_64.rpm: Header V3 DSA signature: NOKEY, key ID f388ca22
Preparing...                ########################################### [100%]
   1:netperf                ########################################### [100%]

Start netperf on any free port number

# netserver -p 16604

Starting netserver with host 'IN(6)ADDR_ANY' port '16604' and family AF_UNSPEC

# netstat -ntlp | grep 16604
tcp        0      0 :::16604        :::*                  LISTEN      3624/netserver

So our port has been opened by netperf

 

On Client

client #  rpm -Uvh /tmp/netperf-2.6.0-3.3.x86_64.rpm
warning: /tmp/netperf-2.6.0-3.3.x86_64.rpm: Header V3 DSA signature: NOKEY, key ID f388ca22
Preparing...                ########################################### [100%]
   1:netperf                ########################################### [100%]

In the below example we are sending traffic for 100 seconds to host server IP 192.169.32.15

# netperf -H 192.169.32.15 -p 16604 -l 100
MIGRATED TCP STREAM TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to 192.169.32.15 () port 0 AF_INET : demo
enable_enobufs failed: setsockopt
Recv   Send    Send
Socket Socket  Message  Elapsed
Size   Size    Size     Time     Throughput
bytes  bytes   bytes    secs.    10^6bits/sec

11796480 11796480 11796480    100.07    994.23

So bond0 network traffic is within our limit as provided to the interface i.e. 1 GB/s

Lets validate bond1

Here my server IP of bond1 is 192.169.32.37

# netperf -H 192.169.32.37 -p 16604 -l 100
MIGRATED TCP STREAM TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to 192.169.32.37 () port 0 AF_INET : demo
enable_enobufs failed: setsockopt
Recv   Send    Send
Socket Socket  Message  Elapsed
Size   Size    Size     Time     Throughput
bytes  bytes   bytes    secs.    10^6bits/sec

11796480 11796480 11796480    100.02   2982.72

bond1 network traffic is also within our provided limit i.e. 3 GB/s

I hope the article was helpful

Leave a Comment