what is the best way to identify network i/o bottlenecks

what is the best way to identify network i/o bottlenecks
we are suspecting there may be a network bottlenecks between our application and database servers. what is the best tool to confirm the bottleneck
Who is Participating?

[Webinar] Streamline your web hosting managementRegister Today

NopiusConnect With a Mentor Commented:
The simplest tool to monitor network load is 'iptraf' and 'vnstat'. The first will give you 'live' statistics on network bendwidth, currently used, the second will collect the same statistics and can show you 'hourly' 'monthly' and 'weekly' traffic load. Both tools are console based.

I'm attaching screenshots of iptraf (live) and vnstat (collected) statistics.

OPNET tools  (www.opnet.com)

OPNET Ace can pinpoint pretty much any network or application response issue.
madunixConnect With a Mentor Chief Information Security Officer Commented:
install MRTG to monitor network devices and servers http://oss.oetiker.ch/mrtg/ 
The new generation of project management tools

With monday.com’s project management tool, you can see what everyone on your team is working in a single glance. Its intuitive dashboards are customizable, so you can create systems that work for you.

balasundaram_sConnect With a Mentor Commented:
App & DB servers are in different networks, communicating via multiple routers/firewall.  How did you confirm network bottleneck,  do you see a lot of DROPPED network packets?
First step would be running TCP trace on each server and enable network spoofing and monitor the traffic.
Note, this monitoring could also impact the performance.
NopiusConnect With a Mentor Commented:
Here is vnstat hourly view.
madunixConnect With a Mentor Chief Information Security Officer Commented:
i use alos nmon for my AIX and linux server to find out more about performance network/cpu/disk beside top/vmstat/iostat/netstat.
The nmon tool is designed for AIX and Linux performance specialists to use for monitoring and analyzing performance data, including:
CPU utilization
 Memory use
 Kernel statistics and run queue information
 Disks I/O rates, transfers, and read/write ratios
 Free space on file systems
 Disk adapters
 Network I/O rates, transfers, and read/write ratios
 Top processors
 IBM HTTP Web cache
 User-defined disk groups
 Machine details and resources

Generally in network adapter: if delay in your network  communications is detected (CRC errors, too many collisions, and so on) then overall performance of client-server communications is affected
So you need to give another level check the network speed settings of the NIC(network adapter settings)..check the switch port settings if you have access, also via netstat on the NIC ( man netstat) check CRC/Collision errors.

You can also use something like OpenNMS. It takes more work to setup but it will stat your entire environment. Not to mention the alerting and trending, etc.
madunixConnect With a Mentor Chief Information Security Officer Commented:
beside above i use open-source pack ......Cacti , MRTG  and nagios to monitor
look @:
http://www.groundworkopensource.com GroundWork Monitor Enterprise
http://www.hyperic.com Hyperic HQ
http://www.icinga.org Icinga
http://oss.oetiker.ch/mrtg/  mrtg
http://www.nagios.com Nagios
http://www.n-able.com N-central
http://www.op5.com op5 Monitor
http://www.open-itcockpit.com openITCockpit
http://www.opennms.org OpenNMS
http://www.hp.com/de/bto HP OpenView
http://www.bmc.com BMC PATROL ProactiveNet Performance Management
http://www.shinken-monitoring.org Shinken
http://www.ibm.com/software/de/tivoli IBM Tivoli
http://www.ca.com/de CA Unicenter
http://www.whatsupgold.com WhatsUp Gold
http://www.zabbix.com Zabbix
http://www.zenoss.com Zenoss Enterprise
All Courses

From novice to tech pro — start learning today.