3 Replies Latest reply on Jan 19, 2011 6:10 PM by tjbradford

    Switchport Receive / Transmit Utilization - Solarwinds Perspective

    myounger

      Can someone please clarify for me what RECEIVE and TRANSMIT mean when looking at utilization on a switchport? Sometimes I see more than 100% on the TRANSMIT for our WAN edge router connected to our 6509.

       

      Does RECEIVE = in to the switchport from the connected device or from the switch backplane into the switchport?

      Does TRANSMIT = out of the switchport to the connected device or from the switchport to the switch backplane?

       

      Thank you.

        • Re: Switchport Receive / Transmit Utilization - Solarwinds Perspective
          pyro13g

          Has nothing to do with back plane.  Directly in and out of port.

          • Re: Switchport Receive / Transmit Utilization - Solarwinds Perspective
            sean.martinez

            He is correct, the Transmit and Receive is for the Traffic on that port only.

             

            This is the same calculation Orion uses to report Bandwisth Utilization for interfaces as well for reference:

            How To Calculate Bandwidth Utilization Using SNMP

             

            If your Cisco Device has 20MB Interfaces and above, I usually recommend to enable 64 bit counters. For higher utilization interfaces such as 100MB and 1Gb Interfaces, it will take only a few minutes for the 32 bit counters to reach their maximum before the counter resets.

             

            To enable 64 bit counters go into Node Management, edit the Device, and under SNMP select "Allow 64 bit Counters" and select Submit. The next polling cycle, Orion will attempt to see if the device is 64 bit capable, and will get the corresponding 64bit counter MIBs.

              • Re: Switchport Receive / Transmit Utilization - Solarwinds Perspective

                worth checking the 64 bit counter are supported first ? 

                 

                also the counters rolling over isn't a massive issue unless your poller is set to only look after extended periods of time 

                there are two methods for dealing with counters rolling over 

                 

                method 1 

                is a difference sum, looks at what there was and what there is now after the last poll and graphs accordingly

                method 2

                same as before but if the counter rolls over it will pause for 20 seconds and check the sum again and graph according to that

                 

                if your network is busy and your seeing the odd spike in your graph every so many mins then is due to it being set to method one and the 32 bit bit counter rolling over, as stated above 64bit counters are the best option but if thats not possible (limitation of device) then method two might be better ? 

                 

                 

                BTW RX traffic is heading into the switch from external 

                TX Traffic is escaping out of the switch router or whatever the device is, when mapping diagrams ensure you dont confuse yourself by applying tx & rx from one switch to the interface of another as this will be the wrong way around.

                 

                " please correct me if i'm wrong "