I was asked a question today about how Orion's Universal Device Poller (UnDP)
works within a large, multi-poller environment. Usually I try to keep these blog
entries somewhat product agnostic but this is a question I've been asked a lot
and that a lot of you may encounter so I thought I'd spend a few minutes
explaining.

 

First off, let's setup the environment. In the scenario we're talking about
you've got a main Orion server and because of the number of devices that need to
be managed you've also deployed two Orion Additional Polling Engines. As you
know, Orion consists of several main components including the web UI (where most
of your interfaction with Orion takes place), the SQL database (where the data
is stored), and the "polling engine" which is a collection of processes and
services that do the work of communicating with network devices to provide fault
and performance management information.

 

Now, let's say that you've went out and bought some new WAN accellerators
from company "BNHBGY" which stands for "Brand New and Hasn't Been Gobbled up
Yet". These devices are brand spanking new on the market but are really cool so
you've depoyed them all over your network (in areas managed by each of your 3
Orion servers mentioned above). You want to be able to measure how the devices
are performing so you've decided to use Orion's UnDP to create custom pollers
for them.

 

You would go to the main Orion server and create these custom pollers. From
there, you can assign the poller to all of the devices of brand "BNHBGY". Orion
will automagically figure out which Orion server (the Orion SLX Server or one of
the two additional pollers) is managing the devices and handle things from there
- no need to do anything on the additional polling engines.

 

For those of you that have been using Orion for a while this might be
surprising as in the old days some of these things had to be done on the
individual servers that were managing those devices. That's one of the cool new
changes that Joel Dolisy (our Chief Architect) and all of the guys here on our
engineering team have been working on.

 

Also, for those of you that are new to the process of creating and/or sharing
UnDPs - keep in mind that the easiest way to do this is to go to the Content Exchange area on Thwack and download UnDPs that other users have
created and shared. I very rarely have to create one myself anymore as so many
users are contributing to Thwack nowadays.

 

Anyhow, that's it for today. I've been in the air all week and am about to
board another plane. Hope you've all made plans to come see us at Cisco
Networkers Live this year - it should be a blast.