Are there guidelines available for use cases where multiple DataMiner systems run separately instead of being combined into a single large DataMiner cluster (e.g. DataMiner Systems provided by Skyline, other DataMiner Systems provided by OEMs like Ateme, Sony, Comtech EF Data or systems that need to run different software versions and can therefore not be combined, etc.)?
Say for example that you want to build a top-level DataMiner cluster as an umbrella system that retrieves and aggregates the most critical alarms from all other DataMiner Systems. That top-level DataMiner System will be connected to OSS/BSS tools like Planned Maintenance, Ticketing, etc.
Which is the recommended way, e.g. forwarding alarms via SNMP, element replication, DataMiner data sharing, etc. to build such an architecture?
For your example, SNMP alarm forwarding is the best option, as you mention that the different systems have different software versions.
Element replication requires all DMS's to run the same software version, as explained here.
DataMiner data sharing is a new feature and at the time of writing will only work on the most recent versions, so again it is not possible to use this option with different software versions.
For other use cases, element replication or data sharing can be good options, so it is best to evaluate this case-by-case.