Efficient Network Access for Autonomous Devices During Times of Congestion
Determines why the machine can’t access the network and takes action based on the cause.
Stagger It! Use Case
Problem: Network congestion risks due to autonomous machines being inefficiently programmed for network access, overloading networks in situations such as an earthquake, causing any device with a motion detector to try to report in immediately and continuously.
- SIMbae enables the IoT Service Provider to provide the app network access software on the SIM, with a simple API for developers integrating their app to IoT Service Provider defined Enhanced Back-Off Algorithm
- Conventional Back-Off Algorithm: increasing delay per attempt
- Enhanced Back-Off Algorithm: delay depends on signal strength, number of attempts and random time offset
- Random time reduces likelihood of devices re-trying in “lock step” in response to an external trigger event (e.g. power outage reporting)
- Signal strength adjuster – low signal strength can cause connection failures that can be mistaken for network congestion
A possible real world scenario:
- 10s or 100s of thousands of machines have accelerometers and motion detectors
- There is an earthquake and the apps in those machines detect the reportable event and try to report in via the mobile network
- Via the random delay back off mechanism, SIMbae Service Assurance Back-Off Timer ensures the staggering of the machines’ network access attempts
- After a specified number of unsuccessful attempts, the Service Assurance Back-Off Timer will assess the signal strength to determine if that is the cause of the inability to connect, and take predefined action to increase the chances of success
Once SIMbae is in place, IoT Service Providers can easily add the Service Assurance Back-Off Timer module. This value added module enables the IoT Service Provider to provide network access software via a standard API and allows the IoT Service Provider to control how and when devices try to access the network.
For example, the network access software can include an algorithm that uses a random number to determine how many seconds a machine will wait before trying to access the network after a reportable event is detected. This method staggers the connectivity attempts of multiple machines, avoiding potential cell site and network congestion. It also has intelligence that determines why the machine can’t access the network (i.e signal strength or congestion) and takes different actions based on the cause.