Hi Dojo, just a follow-up from this original question, as I haven't retrieved all the info I need in terms of NATS.
What's the reason why NATS became "required" from DM 10.2?
Is it contributing in any way to the cluster synchronization or is that still based in full on SLNet?
In need of some clarifications due to specific conditions encountered on specific DMAs
(e.g. intra-WAN tolerated latency &/or super-cluster/leaf-nodes support in 10.4 ).
Thanks
P.S.: Occasionally, I've spotted this kind of stack-trace,
where an attempt is done to send an SLDataGateway message over NATs - any steer will be helpful
Sending SLDataGateway message request:DataRequest<Alarm> with filter:
((Alarm.DataMinerID[Int32] ==*****) AND (Alarm.RootAlarmID[Int32] ==116025154))
over NATS failed with error: DataMinerMessageBroker.API
![](https://community.dataminer.services/wp-content/uploads/wp-user-manager-uploads/2020/10/nomnom.jpg)
Hi Alberto,
We chose NATS because it provides a much more efficient solution to distributed messaging than the communication systems we had at the time. New features use NATS when it makes sense to use then. Old features, like cluster synchronization, keep using the directed communication streams such as via SLNet. When time and priorities allow, we can make efforts to refactor these older features to make use of NATS when it makes sense to use this type of communication.
Hi Alberto,
For the error you mentioned above, would you be able to post the full error? Otherwise, it’s difficult to tell what exactly this is caused by.