This discussion has been locked. The information referenced herein may be inaccurate due to age, software updates, or external references.
You can no longer post new replies to this discussion. If you have a similar question you can start a new discussion in this forum.

Polling Engine Question

Can someone explain polling engines and the number of them? Why would a company have 9??

I've only ever seen setups with single engines and single instances of NPM deployed.

  • I personally have used polling engines for two different use cases.

    The first use case is to add additional polling capacity.  In this case you should be maxing out the capacity of your entire Orion environment long before you max out the capacity of 9 polling engines.

    The second use case is for different security zones.  I have used additional polling engines to contain all of my polling in a specified security zone as to keep from opening too many ports through the firewall for different types of monitoring.

    I hope this helps!

  • Very much so. Thanks a ton.

    SW

  • I'm specing out a solution for <>3,000 devices. Can I ask how you go about deciding the number of polling engines required?

  • A polling engine can typically handle 10,000 monitored elements (elements will be a combination of nodes, interfaces, VLAN's etc). I guess NPM is expected to get a bump up on this 10,000 number in the next edition.

    One straightforward strategy to use is 3000 devices cater to lets say 20,000 users - which directly means that there are 20,000 ethernet ports or access ports. Add some server ports, then add the WAN links, add backbone ports, add security device port counts and add some spare capacity ports.

    Also look at geographical spread - look at your device density in different locations. Polling engines per geographic location to cut down on management traffic through expensive WAN circuits.

    Hope this helps.

  • Thanks RaviK.

    Never thought about geographical dispersal. Something that would be useful in my environment.

  • APE's provide scalability for the Polling Engine piece of the application.  Orion is comprised of many working parts, some of which have limitations, especially considering this is an active polling solution.

    Scalability comes in the form of the type of elements you are monitoring.  I've seen a single instance of Orion with 14 pollers, however, 80 to 90% of the elements were component monitors with SAM.

    Limitations come in the form of how much incoming data the database can handle.  An instance with 9 to 14 pollers reaches the limit of that incoming data to the database.

  • I'm using the polling engine.I will try to explain to you my need for you to understand the rationale for the use of polling engine.

    My company is a provider of IT services. So we have many customers and most of them are not directly connected to our network via VPN or any other similar way. Thus, the only way to access the world of my client is via the internet, via IP's valid. The polling engine works for me in such cases, I have my Orion servers in my company and installed the polling engine in the company of my clients to collect data locally and send me via the web.

    The solution to expensive polling engine leaves you or does not seem feasible, the alternative would be the use of EOC tool. Assuming my scenario, with her I would install all Orion servers in my clients and each would have its local database and independent as I would have the EOC in my business seeking and centralizing the data in the Orion interface that only I have access to view.

    I hope this helps.