7 Replies Latest reply on Jan 23, 2017 5:10 PM by designerfx

    Simple Display of Available Bandwidth

    neovolt

      Does NTA have a function, like iPerf, to capture data on available (unused) bandwidth between 2 interfaces? The reason being that we have buildings distributed over a 3rd party microwave network, and sometimes a 100 meg connection will degrade to 50 meg and it would be awesome if NTA could alert me of that.

       

      Also, I'd love to hear recommendations on what reports are the most useful on interfaces for just general performance monitoring.

       

      Thanks!

        • Re: Simple Display of Available Bandwidth
          mesverrum

          NTA does not have that kind of bandwidth, the netflow/ipfix protocols don't have any fields that I've seen to list that info so NTA can only work with what it is given. 

           

          Does NPM display the correct bandwidth for those microwave links or does it just always show their "rated" speed?  I don't have one handy to test with but I wouldn't be surprised if there was an SNMP OID that would be able to tell you what the current data rate is, you should try an SNMP walk against the device and see if you can hunt it down.

           

          -Marc Netterfield

              Loop1 Systems: SolarWinds Training and Professional Services

          • Re: Simple Display of Available Bandwidth
            jdjones1992

            Neovolt,

             

            If i am reading your question correctly..you should be able to setup an alert with the trigger condition set to the interface bandwidth. That way, if one of the interfaces then cut in half you would receive an alert.

             

            JJ

            2 of 2 people found this helpful
            • Re: Simple Display of Available Bandwidth
              designerfx

              I'd wonder if the standard netflow bandwidth utilization component (on a view/page) might show it? Does the Top 10 NetFlow Sources by % Utilization chart actually update based on interface bandwidth and therefore said chart would show 30mb of bandwidth going from 30% utilization to 60% on the live chart component? Curious, cobrien ?

                • Re: Simple Display of Available Bandwidth
                  cobrien

                  Short answer: no

                   

                  Original poster is asking for available bandwidth.  There are two methods for deriving this today:

                  1. Flood the pipe and see how much fits.  This is what Iperf does, but it often impacts transit traffic because it causes bandwidth contention.  Because of the impact, it's generally not suitable for a monitoring system because monitoring should not introduce new production problems.  You could use SolarWinds' WAN killer to do this today, but I'd say iPerf is a better choice for that use case if you can figure it out.
                  2. Estimate using some fancy probing.  This doesn't introduce performance problems, but has a lot of limitations.  SolarWinds tools do not currently use this method anywhere.
                  1 of 1 people found this helpful
                • Re: Simple Display of Available Bandwidth
                  rschroeder

                  Go with cobrien's answer--it's the best one you'll find.

                   

                  NTA doesn't measure what's not used, only what's used.  And even that measurement must be well understood to be useful.  NTA won't always give you the aggregate bandwidth used in both directions--usually it's only telling you inbound or outbound traffic on a port.

                   

                  To be able to alert on lost bandwidth, Orion would need to know about the unused bandwidth, and then know about a change in it.  You'd need to be sporadically running the equivalent of WAN Killer on that link to determine what's available.  Then you'd need to compare that data against what was running across the link pre & post probe period, and find the delta (if any) between future and past probes.

                   

                  Although it's not the answer you wish for, NTA doesn't prove negatives--it won't tell you what unused bandwidth is no longer available.

                   

                  You might be able to do some NPM alerting on dropped packets or resends, though.  If your 100 Mb microwave pipe were 90% utilized and then the microwave circuit throughput capacity dropped to 50 Mb, you'd have retransmits and / or dropped packets until the problem resolved.  NPM would record those numbers increasing, and you might be able to create an alert based on retransmits or discards increasing on those ports, but it's not the information you're looking for--it's more of a measurement of the symptom or consequences of the microwave issues.

                   

                  Certainly you'd wish for QoS to be in play when your microwave pipe drops from 100 Mb to 50 Mb, but QoS would need some kind of "available bandwidth discovery tool" working to set it into motion, and that's just not out there.  QoS is just static thresholds, not dynamic discovery of available bandwidth that automatically adjusts QoS settings.

                   

                  I've heard weather can cause microwave data throughput reduction.  I keep a NOAA Weather Radar real-time view in a widget/window in NPM's front page to help me know when there are thunderstorms in my regions; they're a significant percentage of the causes of WAN outages as they take out power to cities and substations.  You might consider that option, too, if heavy rain or snow is dropping throughput capabilities.  It won't alert you, but will help you put 2 & 2 together when you see bad weather around the microwave areas and you start getting WAN complaints or performance issues.

                   

                  I suspect you're using microwave because direct fiber feeds aren't available; maybe you're in very mountainous areas, and wireless is the only affordable solution?  If not, search for a vendor that will provide wired throughput and you're microwave issues can be moved to a secondary backup option.

                   

                  While thinking of performance issues, if you're using the Quality Of Experience tool in NPM, you could possibly alert on increased issues and latency of applications there, too, if they pass across those microwave towers.

                   

                  If you come up with a solution, please post it here for others to learn from.

                   

                  Swift Packets,

                   

                  Rick Schroeder