Hi There,

During my time in corporate IT, we had to alert receivers of alerts to the
fact, that they might get a false positive. This was happening quite often
from remote sites. Dependencies did remove a lot of alarms caused by
downstream devices. However, false positives still remained.

Digging a bit further into the issue revealed a common scenario for larger
frame relay based networks. In a typical corporate environment with mostly
bursty traffic, you will see any given link (unless it is a fat pipe) being
saturated on and off for a short while one way or another. This means,
traffic above a certain threshold gets flagged for discard eligible, means
it can be dropped without notice. ICMP traffic is not very high on the
priority list anyway. Many times, carriers will drop ICMP traffic in favor
of "real" data traffic adding to the issue.

One could alter the polling engine such a way, that ICMP pings will get
issued more then once if a single miss. This will of cause mean a redesign
of the polling system as this could lengthen the polling cycle beyond the
map cycle. Perhaps a special repolling queue ore something similar would do
the trick.

This would also enhance the value of dependencies. Downstream sites who
receive alarms very often do not know the state of the central routing
system. It would be beneficial if this can be placed on the submap without
the need of a complete new device object. It should be possible to call an
objects interfaces programmatically regardless of the current polling cycle.
If this sounds a bit to technical, it should be possible to place an
existent device on a submap.

I hope this is not to complexly worded. In short, handling of dropped ICMP
packets and reuse of map items on several maps.


Luz Berger
Berger Network Consult
http://www.bergerl.com


-----Original Message-----
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Adrian Ferrier
Sent: Saturday, May 29, 2004 9:43 PM
To: [EMAIL PROTECTED]
Subject: [WhatsUp Forum] User Stories on Dependencies - research request


Hey All,

 I'm interested in gathering end user stories on how you are using Up (and
Down) Dependencies on your maps.  What is the basic value you gain from this
feature?  (don't worry it's not going away)

 We already know that there is a desire to be able to create dependencies
that span maps (although feel free to discuss further). And we already know
that it is a challenge to resolve invalid dependencies through manipulating
poll order.  Those things aside - how are you using them? - what's your
polling dependency story?

 Feel free to respond to the list if you would like others to view and
comment or feel free to just respond directly to me.

Regards,
Adrian Ferrier
aferrier at alpha.ipswitch.com
WhatsUp Development Manager, R&D
Ipswitch, Inc.


Please visit http://www.ipswitch.com/support/mailing-lists.html 
to be removed from this list.

An Archive of this list is available at:
http://www.mail-archive.com/whatsup_forum%40list.ipswitch.com/


Please visit http://www.ipswitch.com/support/mailing-lists.html
to be removed from this list.

An Archive of this list is available at:
http://www.mail-archive.com/whatsup_forum%40list.ipswitch.com/

Reply via email to