During a polling cycle, a connector thread determines the full set of active measurement points for that connector and splits it into smaller subsets. Then, subset by subset, it determines the latest readings for the measurement points in the subset and updates the measurement points accordingly.

As described, a thread handles the subsets of measurement points sequentially. The bigger the total number of active measurement points, the higher the number of subsets, and the longer it takes to process these subsets. Since most polling cycles are configured to poll every minute, there is a maximum number of active measurement points that a connector can handle within that minute.

To scale, it is possible to have multiple Planon IoT connectors (threads) running in parallel. The benefit of parallel execution is that it can update multiple subsets of measurement points at the same time. Since one Planon IoT connector only uses a small part of the memory, CPU and storage capacity it is possible and recommended to configure multiple connectors. This allows for more flexibility (each connector can have its own configuration) and in general leads to a better scaled system with the best performance.