Re: [nv-l] Bandwidth usage Benchamrk

[James Shanks <jshanks@us.ibm.com>]

First, I have no benchmarks, though
you can open a problem to NetView Support and see whether they have anything
they can share with you.  Or you can look for white papers and other
sources out on http://www-306.ibm.com/software/tivoli/products/netview/

The only tuning for polling you can
do is xnmsnmpconf.  The rule of thumb is don't set your polling intervals
so short and your retries so high that they overlap.  You determine
the collection rate and the amount of data collected for Data Collection
yourself.  Again the rule of thumb is don't ask for too much too quickly.
 And I would observe that the amount of bandwidth for any of this
depends upon your network and what you ask for.  So you could try
it and get a benchmark of your own with snoop or an iptrace in a very short
time.

Finally, NetView does not waste bandwidth
unless you tell him to .   Use a seed file to restrict discovery to
just those nodes you want to manage or unmanage them once the unimportant
ones have been discovered.   Don't collect data unless you really
intend to do something with it.   And use RFI.  It reduces pointless
polling by netmon when a network has become occluded.

More than likely you will waste a lot
more bandwidth than NetView ever will by configuring your routers and switches
to send pointless traps to NetView every other second.  That is the
biggest waste of bandwidth in most shops.  The best thing you can
do, in my opinion, is to start a  policy that no trap is to be sent
to NetView unless someone at the management station is going to actually
do something with it when it arrives.  That's how you can prevent
a trap storm and save your bandwidth.


James Shanks
Level 3 Support  for Tivoli NetView for UNIX and Windows
Tivoli Software / IBM Software Group

Meyos Yemveng <yemveng@yahoo.ca>
Sent by: owner-nv-l@lists.us.ibm.com

03/02/2004 07:52 PM

Please respond to
nv-l

To	netview Tivoli <nv-l@lists.us.ibm.com>
cc
Subject	[nv-l] Bandwidth usage Benchamrk

Hi Folks,
Q1: Anybody has any benchmarks on the following:
1-       Average
Bandwidth usage by NV for managing 1000 Nodes
2-       Average
Bandwidth usage for a Data Collection at a 15 min interval
3-       Average
Bandwidth usage for status polling  
4-       Average
Bandwidth usage for initial discovery
5-       Average
Bandwidth usage for rediscovery
 
Q2: Any Hint or advice on how to
minimize or prevent wastage of Bandwidth by netview?
 
Really appreciated.

Thank You,
Meyos
"ALLEZ LES LIONS INDOMPTABLES"

---

Post your free ad now! Yahoo!
Canada Personals