topic RE: Peak Tx & Rx % bandwidth in ExtremeSwitching (EXOS/Switch Engine)

Peak Tx & Rx % bandwidth

EtherNation_Use — Wed, 08 Jan 2014 05:56:00 GMT

Create Date: Oct 10 2012 7:00AM

Hello
How does XOS calculate Peak bandwidth to the command "show port util"?

We sometimes have big differences in values from our switches and in Cacti, and I'm trying to figure out why..

BR
Lelle (from Lennart_Berglund)

RE: Peak Tx & Rx % bandwidth

EtherNation_Use — Wed, 08 Jan 2014 05:56:00 GMT

Create Date: Oct 10 2012 12:06PM

Hi Lelle,

Keep in mind that numbers shown in 'show port util' are sampled data. I believe the sampling frequency is something like 3-5 seconds. So numbers can vary drastically.

I would trust the information Cacti gathers via SNMP much more so than anything shown via telnet/ssh/console.

Thanks,
Andrew (from Andrew_McConachie)

RE: Peak Tx & Rx % bandwidth

EtherNation_Use — Wed, 08 Jan 2014 05:56:00 GMT

Create Date: Nov 15 2012 4:41PM

In my experience, cacti and MRTG have underreported utilization more than the console session has. I didn't build the cacti config so I can't vouch for its correctness, but on a bursty link that carries a true average of 200Mbps shows up as about 50Mbps in cacti/MRTG no matter what, and 150 at the console unless it's sustained (the longer it's sustained at a given speed the more correct the console gets.) Each of these tools is showing data that's sampled then averaged, like the poster above said, so it's a matter of how exact you want to get. If you need the most accurate numbers, you need to sample more frequently. (from Ansley_Barnes)