It's a variety of products, each separately licenced
a) Additional polling engine -- used to increase the element polling capacity
b) Additional Web server engine -- used to offload the webUI
c) Failover engine for DR/HA environments
I already hold this document, so as per your answer I confirm that scalability engines are a variety of product.
So the new question will be : how to automatically load balancing the APEs?? if this cannot be some things embded in the APE or Orion platfomr I think the only way (wich will be a bit hard) is to script some things that will check the Polling Engine healt and performance and based on the result execute an API script that will move the carried nodes to an other polling engine!!
If it's the case, could you please provide some guidences.
Thanks and Regards.
I don't bother balancing across APE -- there is not much point, and it does not get you anything (assume they both have about the same power and capable of supporting the full load).
In my environment we have almost nothing on the primary polling engine -- that server is dedicated the the WEBui, syslog, traps, and Netflow data capture.
New nodes are added to the primary server, and periodically they are moved to the least loaded additional polling engine manually.
Here is my current load (by interfaces)
Periodically I select the nodes on the app1 server, and move them to the least loaded server(poll3 right now).
I'm fairly skilled in using the API, and the juice ain't worth the squeeze to do this automatically.
In some environments the different APE might have different network connectivity, in which case you would need to manage the location of nodes using some other rules, and the whole situation becomes quickly out of hand.
you also have to add hysteresis (where moving nodes around causes some imbalance, which causes you to move nodes around, etc... by doing this manually we have a much more stable environment)