Skip to Main Content

Hardware

Announcement

For appeals, questions and feedback about Oracle Forums, please email oracle-forums-moderators_us@oracle.com. Technical questions should be asked in the appropriate category. Thank you!

Measuring Sub-second Network Bandwidth

3311079Sep 9 2016 — edited Sep 12 2016

Hi,

Does the community have any suggestions on measuring network bandwidth at sub-second granularity, which do not require root access?
We have tried collectl, which does support sub-second intervals.  However, only 1 line per second of the output contains bandwidth data.  The rest are zeros (see below).  We have observed this on both 1/10 and 1/100 second intervals.

The issue we are dealing with is brief bursts data coming in at very high rates.  If we have a 1 gig circuit, are there sub-second intervals where bandwidth exceeds 1 gig, even though over the whole second, bandwidth is less than 1 gig?

Packet captures have proven this to be true, however these require root access which is difficult to get in our environment.  Also, packet captures are useful only after the fact.  We need a tool or technique that would provide basically real time reporting.

Thanks in advance for all insights!

################################################################################
# Collectl:   V4.0.5-1  HiRes: 1  Options: -sN -oDm -i.1 --netfilt eth1 -P -f col10
# Host:       bosa01  DaemonOpts:
# Booted:     1428927446.14 [20150413-08:17:26]
# Distro:     Red Hat Enterprise Linux Server release 6.4 (Santiago)  Platform:
# Date:       20160907-121853  Secs: 1473265133 TZ: -0400
# SubSys:     N Options: Dm Interval: .1 NumCPUs: 4  NumBud: 0 Flags: i
# Filters:    NfsFilt:  EnvFilt:  TcpFilt: ituc
# HZ:         100  Arch: x86_64-linux-thread-multi PageSize: 4096
# Cpu:        GenuineIntel Speed(MHz): 1999.998 Cores: 2  Siblings: 2 Nodes: 1
# Kernel:     2.6.32-358.el6.x86_64  Memory: 1920132 kB  Swap: 4576248 kB
# NumDisks:   1 DiskNames: c0d0
# NumNets:    7 NetNames: lo:?? eth0:100 eth1:1000 eth2:1000 eth3:100 virbr0:?? virbr0-nic:??
# SCSI:       CD:0:00:00:00
################################################################################
#Date Time [NET:eth1]Name [NET:eth1]RxPkt [NET:eth1]TxPkt [NET:eth1]RxKB [NET:eth1]TxKB [NET:eth1]RxErr [NET:eth1]RxDrp [NET:eth1]RxFifo [NET:eth1]RxFra [NET:eth1]RxCmp [NET:eth1]RxMlt [NET:eth1]TxErr [NET:eth1]TxDrp [NET:eth1]TxFifo [NET:eth1]TxColl [NET:eth1]TxCar [NET:eth1]TxCmp [NET:eth1]RxErrs [NET:eth1]TxErrs
20160907 12:18:53.402 eth1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
20160907 12:18:53.501 eth1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
20160907 12:18:53.602 eth1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
20160907 12:18:53.701 eth1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
20160907 12:18:53.802 eth1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
20160907 12:18:53.901 eth1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
20160907 12:18:54.001 eth1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
20160907 12:18:54.101 eth1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
20160907 12:18:54.201 eth1 23830 9170 24161 627 0 0 0 0 0 0 0 0 0 0 0 0 0 0
20160907 12:18:54.301 eth1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
20160907 12:18:54.401 eth1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
20160907 12:18:54.501 eth1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
20160907 12:18:54.601 eth1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
20160907 12:18:54.701 eth1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
20160907 12:18:54.801 eth1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
20160907 12:18:54.901 eth1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
20160907 12:18:55.001 eth1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
20160907 12:18:55.101 eth1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
20160907 12:18:55.201 eth1 58580 18800 70545 1285 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Comments
Post Details
Added on Sep 9 2016
0 comments
424 views