bandwidth monitoring

what do you guys use to monitor you bandwidth? to see how stable you connection is or when it drops things like that

Bandwidth at which point… possible monitoring tools are :

Free:
MRTG
CACTI

Others:
Intermapper
PRIZM
BAM

All monitoring tools (not sure about Prizm) are simply doing SNMP queries to collate the information. When setting up becareful not to create too many SNMP queries it could affect your bandwidth.

We poll our core kit every 10 seconds. For bandwidth and more static stats its about every 20 seconds. Only monitor SM if investigating a problem.

We use Cacti and it works great. Check out CactiEZ (google it).

I am also using CactiEZ. It was free and very “EZ” to set up.
Jerry - how often are you polling your AP’s, BH’s and SM’s and what stats are you collecting?

It was free and very "EZ" to set up.

Got that right. I am not a Linux guy by any stretch of the imagination. This was literally plug and play. Took longer to download the .iso and burn the disk than the actual installation.

One of our Engineers said that Cicso charges a couple grand just for their router monitoring software. This whole package represents about 4000 worth of software and so far has proven to be more stable, easier to use, and more stable than any of the other software we tried (and I think we tried everything with a free 30 day eval.)



Whatever the default polling interval is. Have not needed to change it.

One of the best tools is NTOP. Being able to look at and analyze the types of traffic on the network is huge.

The Device monitoring is great . I like getting an alert on my phone when a node goes down. I just need to remember to turn off Cacti when I am doing maintainence - I had 200 alerts in 4 hours yesterday when I was working on the towers.

We are montoring the following

[u:ijl1dbtp]AP's/BH Masters[/u:ijl1dbtp]
Ping Latency - Advanced Ping Template
Ethernet Link Status - Canopy Template
GPS Link Status - Canopy Template
Registered SM Count - Canopy Template
Temperature - Canopy Template
Traffic - Standard SNMP Template

[u:ijl1dbtp]SM's/BH Slaves[/u:ijl1dbtp]
Ping Latency - Advanced Ping Template
dBm - Canopy Template
Ethernet Link Status - Canopy Template
Jitter - Canopy Template
RSSI - Canopy Template
Session Status - Canopy Template
Temperature - Canopy Template
Traffic - Standard SNMP Template

[u:ijl1dbtp]Router[/u:ijl1dbtp]
Router Astound 3600 - CPU Usage
Router Astound 3600 - Free Memory
Router Astound 3600 - ProcMem Usage
Router Astound 3600 - Traffic - Fa2/0
Router Astound 3600 - Traffic - Fa2/1

can it also tell me how many Gigabyte one client has used in the last month?

i have clients running backup systems that draw up to 60 gigabyte a night. if i would be able to bill those customers by the usage i would be rich. bam is supposed to do it but you need to set up radius for it.
the motorola manual is very, very light on that topic so… i never tried it.

any suggestions? traffic monitorin per ip or something like that?

thanks


gero

Someone set up Cacti for this, but is not sharing the templates.

I would set up a radius server to do it with BAM

Are you using a CISCO at your edge ?

If so I would use cisco cach-flow stats to get this info.

If you anticipate monitoring usage for a significant number of customers, then I would explore the various bandwidth shapers on the market (two birds with a single PO).

If all you need to do to generate a report on a couple of customers (e.g. usage X hour X day) then a small PERL script could pull ifInOctets and ifOutOctets either directly from the Canopy SM via SNMP or from line 1 of a MRTG log file (customer gets a pretty picture as well). The script gets thrown into the CRON to run once an hour and you grab the report when you want to bill the customer. MRTG converts traffic passed into traffic rate when it completes the next polling cycle.

u need linux to run this what ver are you guys running and where can i download it??

Buck,
Check out CactiEZ - download it here: http://www.cactiusers.org.

This is a single disk installation with a stripped version of CentOS Linux with Cacti integrated.

you don’t need to know Linux and as monitoring systems go, it’s very very easy to set up and manage.

Whle PERL & MRTG will run on any platform I use Redhat Linux.
Go to http://www.activestate.com/ for a Windows PERL librry.

for windows there is PRTG i’ve never used it. On Red hat, MRTG comes with the distribution.

FWIW CentOS 4 is RH4 with most of the RedHat branding removed.

Do any of you have an opinion on KS Host Monitor? We have been using this for years to monitor our Proxim 2.4GHz System. I know it is capable of using SNMP and would love to just integrate the 900MHz Canopy into the same monitoring system.

We are rolling over our 2.4GHZ clients to the Canopy 900 to add the NLOS functionality that is desperately needed in the wooded areas of Southern Oregon. The other reason for the change is the dramatic increase in interference of the 2.4 range over the past two years.

I currently have 4 AP’s and 20 SM’s on the Canopy network and am trying to work out the kinks before adding the 100 people on my waiting list. A good monitoring system would be a valuable tool for me.

I am a big fan of mikrotik for such things. We replaced our BAM a couple of years ago due to problems (and resistence to keep paying for a license for each and every AP) and our cisco routers and pix firewalls with them.

Mikrotiks have been doing most everything for us, including bandwidth shaping, P2P and virus blocking, BANDWIDTH graphing (modified MRTG), aggregate client consumption reporting, DNS caching, client traffic prioritization, outage alerts, watchdogs, etc. We got a remarkable stretch for our current bandwidth to client ratios, and the cost is one of the best values in this market. We have MT certified engineers if anyone wants to know more or want us to configure a router for you.

The DUDE software looks nice for monitoring in a Windows system. We use windows. However, with the Microtik or other options outside of Prizm, can you also monitor the Authentication Key to make sure that what you are monitoring is not a Mac - spoof ?
Thanks.