The quality of the cellular connection between a Cloud Connector and DT’s cloud service depends on a number of factors:
- The distance to the nearest base station
- How much radio noise there is in the area
- Network congestion
- Placement and orientation of the Cloud Connector
We strongly believe that a robust connection and good observability are critical for a successful IoT sensor deployment. This is why we have collected tens of thousands of data points from our fleet of 4G Cloud Connectors to assess how we can use the various metrics reported by the cellular modem to improve the signal quality indicator for our customers.
Our signal quality estimation reflects real-world performance and is tailor-made for our 4G Cloud Connectors. We have developed it with inspiration from the Android Open Source Project.
Cellular signal strength between 20–60% is considered normal
Low signal strengths will still be able to deliver a robust cellular connection.
Technical deep dive
The cellular modems used in our 4G Cloud Connectors report several other metrics related to the cellular connection in addition to RSSI (Received Signal Strength Indicator). The three metrics of interest are RSRP, RSRQ, and SINR.
- Reference Signal Received Power (RSRP) is a 4G-specific measurement that represents the average power of cell-specific reference signals. In our 4G Cloud Connectors, RSRP is nearly directly proportional to RSSI.
- Reference Signal Received Quality (RSRQ) aims to describe the quality of the received signal, and takes into account the RSSI, RSRP, and how much of the available bandwidth is used.
- Signal to Interference plus Noise (SINR) is yet another measure of the quality of the cellular connection. In contrast to RSRP and RSRQ, which are defined by the standards organization that writes the technical specifications for cellular networks, SINR is defined by the modem vendors themselves.
For in-depth information about these parameters, you can see ETSI TS 36.214.
A common method of finding the overall signal quality based on these parameters is to compare them with pre-defined limits for “good”, “moderate” and “poor” quality. The parameter with the worst performance determines the final signal quality.
This method works well, but the obvious question is: what limits are appropriate for these parameters? We have taken an empirical approach and looked at the historical performance of our 4G Cloud Connectors. This enabled us to determine the range in which each parameter varies and how the parameters correlate with one another. The result of this is what we believe to be a robust estimation of real-world cellular signal quality.