I'm not sure you can be successful trying to win a political battle with
technical information. If your Network guys don't like NetView and want it
out, and they dispute how you gather and measure its impact, then you are
probably doomed. Why? Because you are essentially duplicating each other.
Each of you monitors the same devices, so even if you both configure your
respective products to the minimum, then you still have twice as much
network traffic than you would with just one manager. I am surprised that
the Network guys have not insisted that you add the OpenView TEC adapter
to their OpenView to send events to TEC. It is harder to configure than
the internal one shipped with NetView but it could be done. As a NetView
guy I don't like the solution, but as a Tivoli guy, it seems reasonable to
settle the issue that way.
But to answer your question objectively, I would add a sniffer. You can
tell precisely how big the pings are how frequent, and the same with the
SNMP config polls.
That should be vendor neutral.
But my opinion is that it is ludicrous of them to maintain that NetView has
any more chance of adding a lot of overhead to the network than OpenView ,
I think, because they both work essentially the same way -- configuration
by SNMP, status by ping. You can control what we do in xnmsnmpconf. I
don't know how they do it. But the only potential for one of them to get
out of control is if someone deliberately messes with the polling parms
after set up, or turns new discovery on, and so on. That could happen to
them as well as to you, depending on how tightly access is controlled.
James Shanks
Team Leader, Level 3 Support
Tivoli NetView for UNIX and NT
Mike Guik <wombat_at_large@yahoo.com>@tkg.com on 01/23/2001 02:52:49 PM
Please respond to IBM NetView Discussion <nv-l@tkg.com>
Sent by: owner-nv-l@tkg.com
To: nv-l@tkg.com
cc:
Subject: [NV-L] Coexistence of NetView and OpenView
Hello everyone,
We are using NetView 6.01 integrated with TEC for
determination whether a heartbeat event is caused by
a host being down or by a router interface being down
(Router Fault Isolation). Our Network support group
uses OpenView and they don't monitor any nodes - only
routers and switches. They are concerned if our use
of NV could lead to a heavy additional load to the
network.
We have restricted the list on servers we monitor -
all of them are listed in seed file and New Node
discovery is turned off.
My question is
1) does anybody use NV along with HP OV;
2) what reliable metrics can be used to estimate
a load generated by NV? I emphasize "reliable" since
all measurement results are subject of debates with
the Network Group.
we can measure package ins/out of server network
interfaces and the Network group collects data on
router interfaces for subnets. Any other suggestions?
Thanks.
Mike
__________________________________________________
Do You Yahoo!?
Yahoo! Auctions - Buy the things you want at great prices.
http://auctions.yahoo.com/
_________________________________________________________________________
NV-L List information and Archives: http://www.tkg.com/nv-l
|