|
Fair enough. Other than the rrdplugin.c & .h, are
there any other files I should be aware of that I'd have to modify? I'm
not a programmer by any stretch, but I can hack other's stuff ok. Any
pointers would be greatly appreciated. Thanks! From:
[EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Burton Strauss That's pretty much EXACTLY what it
does. AS gives you a way to consolidate sparse addresses, without using
sub/supernets or filtering. The downside is that aggregated means you
lose the detail. It may be that what you really need is a
"Dump AS" option in rrdplugin (like "Dump Domains").
That preserves the detail. The code wouldn't be that hard to write w/ the
model from Dump Domains. Have fun! ----- From:
[EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of John Hally So what does doing the AS aggregation in
Ntop buy me? I'm not sure I follow that part. Essentially it would be helpful to be able
to take a timeframe, (say monthly) and be able to see that we passed X-amount
of data to-and-from that AS. I'm looking to do this because we currently
get billed based on usage for traffic to If the above isn't currently available,
I'd definitely be interested in getting involved in moving this to Ntop-dev and
helping out anyway I can. Thanks! From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Burton
Strauss If you aren't going to use the netflow
aggregation, then this becomes an rrd issue. Unfortunately, the summation (aggregation)
scenario is the one that hurts re rrd - you can't just accumulate into an
rrd. One update per second is all it will accept. So, to add RRDs by AS, you would need to
walk the HostTraffic structure, summing in memory and then do one update per
AS. "Dump Domains" is the model - it's not THAT hard to add the
code (but let's move this to ntop-dev if you're interested). ----- From:
[EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of John Hally Thanks What I'm looking to do is to gather usage
stats based on AS. I'm currently aggregating based on AS at the border
routers and doing no aggregation via Ntop, and I think that's working out ok. One question I have is, is there a way to
keep the AS Info stats (RRD?) so that when the process gets restarted I don't
lose what's reported under Summary/AS Info? Thanks! From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Burton
Strauss Basically, it's what YOU need it to be to
meet YOUR needs. Aggregating at the router (netflow
collector, technically) reduces the amount of bandwidth consumed for monitoring
(and reduces the load on ntop). But it's more difficult (sometimes) to
change the collector options. Aggregating in ntop is easy to change
(just flip the dropdown). But it means ntop is processing each flow. Whichever way you aggregate, that's the
level of detail ntop reports. No drill-down, etc. just the
aggregated data. ----- From:
[EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of John Hally Hello All, Can someone tell me what the benefits of the different
netflow aggregation options are? Is it better to not aggregate at the
router, and then set up aggregation on Ntop, or should they 'line up', meaning,
set AS aggregation on the router and also in ntop? I guess which type of aggregation you choose will also
effect the type of reporting that Ntop is going to produce, correct? Thanks in advance! |
_______________________________________________ Ntop mailing list [email protected] http://listgateway.unipi.it/mailman/listinfo/ntop
