Data Optimization & Edge Computing

Data optimization starts with determining exactly what data you need, and how frequently you need it. In terrestrial networks with few data constraints, information is passed freely; when it comes to satellite, consider paring that back.

For example, you could have a simple temperature sensor; the information you’re looking for is the temperature. But quite often it will transmit lots of information about its firmware / software version, its ID, what protocol it’s using, how often it’s transmitting, configuration settings – usually unnecessary, and always costly to send.

Human-readable information is another opportunity to optimize. Frequently – per the above example – we see information being transmitted that includes far more data overhead than is necessary for a computer. Sending the former will cost you more money, and drain your battery faster, than is necessary.

As well as reducing the amount of data you send, you should also consider how often you send data. Could you manage with every 30 minutes instead of every two minutes, for example? Or can you report on exception?

Finally, could you use a messaging service instead of TCP/IP? Messaging is a much more efficient way of using satellite airtime. Read our blog post on this topic to learn more.

Edge computing can help you substantially lower the cost and power drain of your satellite IoT data transmissions. Firstly, you can report on exception: when data falls out of pre-agreed parameters, transmit; otherwise, assume no changes.

Secondly, you can define your data priority; some protocols, notably MQTT, allow you to define different priorities for different data sets. Lower priority data sets can be batched and sent once a day, for example; higher priority data sets sent every 15 minutes.

Thirdly, you can interpret and analyze the data in the field. If a temperature sensor reaches a certain limit, your edge computing-enabled terminal can action a response to this, whether that’s to alert an engineer or start some form of cooling process.

Finally, you can reduce, average or compress your data.

Recently we collaborated on a system for a large wind and solar measurement tool company. Instead of transmitting everything captured by their data loggers, we’re sending a sample of each log file to their server. The server decides what’s worth investigating further, and that log file is then sent in full.

If your sensors or gateway have edge computing capabilities, you can of course utilize them. Equally, we have designed satellite IoT transceivers with in-built edge computing for this specific purpose: to save money and conserve battery power, while still ensuring that you get essential data when you need it.