Network Resource Pool design query

Question

Current Scenario - Two physical sites - Two SCOM Management servers assigned to perform Network Monitoring.
Currently the proposed configuration is to have two separate Network Resource pools so that the network devices in each site is explicitly managed by the network resource pool local to that site. The question is, if one of the sites or network resource
pools becomes unavailable, how do I continue to manage all the devices by failover?
Alternatively, and I may have answered my own question here, if I create one single network resource pool and encompass both SCOM Servers in that pool with an explicit discovery rule assigned to both servers to cover all network devices, that would provide
the resiliency I need, but would the servers on each physical segment still manage the network devices local to them?
Also, I am aware that it is possible to assign an explicit rule to each server within a resource pool. In that scenario, if one of the servers in the resource pool becomes unavailable, does that mean that the network devices monitored by that server are
no longer monitored?
Any advice on which solution to implement would be much appreciated. I understand that there are pro's and cons to all of the scenarios that I have listed. Any links to some more in-depth reading would also be appreciated.
Eddied

Answer 1

First, to your last question, if the network device discovery server goes down, this doesn't affect monitoring at all. The devices will not be undiscovered, and monitoring will continue to happen. The discovery server is completely separate from the monitoring
resource pool. The discovery server does not even need to be a member of the monitoring resource pool at all - it could be some other management server that doesn't actually do any network monitoring.
If you are using true management server (not gateways), then I assume there are no port restrictions communicating to any of the devices in any site or subnet. This means you can simply create a resource pool containing 2 or more management servers, and
do all network monitoring from that resource pool. The resource pool manages which server is actively monitoring which device. I wish I had more information into the algorithm that calculates this under the hood, but I don't know. To address your concern,
though, the network monitoring resource pool is inherently HA - when one MS goes down, load will (or should) be transferred to other members of the pool automatically.
I hope that helps a little more than my previous response.
Jonathan Almquist | SCOMskills, LLC (http://scomskills.com)