Network Management

 View Only
last person joined: 20 hours ago 

Keep an informative eye on your network with HPE Aruba Networking network management solutions
Back to discussions
Expand all | Collapse all

[ASK] How do I monitor IP eating bandwidth on HPE MSR1002-4

This thread has been viewed 0 times

  • 1.  [ASK] How do I monitor IP eating bandwidth on HPE MSR1002-4

    Posted Apr 07, 2019 11:41 PM

    [ASK]

    How do I know which IP addresses consume a lot of bandwidth, routers are configured with DHCP and VLAN interfaces that are used by many users to access applications to servers and the internet?

     




    #traffic
    #monitoring
    #MSR1002-4


  • 2.  RE: [ASK] How do I monitor IP eating bandwidth on HPE MSR1002-4

    Posted Apr 08, 2019 05:19 AM

    Hello,

    If you are on this forum, I assume that you are using IMC.

    You should then look at the NTA module which allows you to exploit both, sFlow and NetSteam sources available in the MSR1002 router. This will give you a lot of inside on "who is doing what" on your network and WAN connectivity.

    Ray

     

     

     



  • 3.  RE: [ASK] How do I monitor IP eating bandwidth on HPE MSR1002-4

    Posted Apr 09, 2019 05:36 AM

    yes im use IMC,

    tks for sugestion, my device support sflow adn will be configure.

    how much bandwidth is needed for sflow, I'm afraid that if installed on a 500 router the client meets the backhoul traffic



  • 4.  RE: [ASK] How do I monitor IP eating bandwidth on HPE MSR1002-4

    Posted Apr 11, 2019 04:55 AM

    sFlow is a very light protocol based on sampling. In short, it picks 1 packet out of 100 (default:configurable), remove the content just keeping layer 1 to 7 header information, add RMON interface statistics, and send the the short resulting packet like a SNMP trap to an aggregator (in your case the IMC/NTA server). So it uses very low bandwith and very little resource on your switch or router, the sampling being done in the ASIC.

    However the IMC server must to handle these samples, aggregate them and load them in the NTA DB in an almost real-time. THIS takes some resources.

    If I understand corectly your environment you want to aggregate the samples of 500 routers. I would not higher than 20 samples/s, which means that you should configure your routers to send 1 sample/30s. This will give you a rough accuracy of bandwidth usage and users.

    The best is to go for a IMC distributed environment and have separate server(s) just for NTA. You could then short the sampling interval to 1/s.

    I hope it helps