Why the traffic difference between NA And Cacti

[eric.lin826]

Hi Guys
As subject,when I monitor the same source, but there is difference result between Nagios Network Analyzer and Cacti, Why?

[tmcdonald]

They use two different protocols. NNA uses netflow, which is sent typically by switches/routers and is basically a summary of the data. Cacti I believe uses SNMP, which polls the device directly for bandwidth usage. If you have two devices sending netflow data for the same network segment, you can end up seeing double usage. Cacti also appears to be showing bits per second, whereas NNA is showing bytes, and it is not clear what Cacti is monitoring (a single port, combined bandwidth for a whole switch, etc.)

[joeywu]

Hi ,

As tmcdonald says , NNA and Cacti are using different protocols and bandwidth units.

Here is something interesting ,i had upload 4 screenshot in attachment , two from Cacti(named start with MRTG) and two from NNA(named starts with VPN).

It seems like NNA traffic is similar with Cacti , and both them are using bits per second .

[joeywu]

Because i can only upload 3 attachment once , so , i post other 2 attachment here

[bwallace]

Thanks for the input joeywu. OP, are there any other questions?

[joeywu]

Hi,

I think maybe we only have these questions in this case, maybe we can discuss why cacti's Graph similar with NNA.

[Box293]

Can you re-phrase your question please? It's not clear what your problem is.

[joeywu]

Hi all

My question is why the graph of bandwidth between NNA and Cacti are different (you can see the attachment which named "different.png"), as tmcdonald says , NNA and Cacti are using different protocols and bandwidth units.

Here is something interesting ,you can see that attachment in attachment (VPNR1-NNA&Cacti.png & VPNR2-NNA&Cacti.png)

It seems like NNA traffic is similar with Cacti , and both them are using bits per second .

So , my question is what can make it different ?

[tmcdonald]

Without knowing more about how your sources are set up in NNA compared to what you are monitoring in Cacti, it will be very hard to determine why they are showing sometimes different and sometimes similar results. My assumption is that when the results are different you are getting traffic in both directions (or from multiple sources) in NNA, as opposed to just a single source (SNMP port) in Cacti. When they are similar, you're possibly getting traffic in both directions from both products.

[joeywu]

For NNA , we are setting 3 netflow to from 3 different device , and all of our compare source (NNA & Cacti) are the same.

And every our netflow source is from single device single port.

So , if NNA and cacti using different unit of traffic , why their bandwidth graph looks the same ?