Edge computing with an “Internet of Things” based sensor array : an innovative approach to near real time seismic exploration and monitoring.


We demonstrate the feasibility of leveraging an Internet of Things (IoT)-based sensor array to orchestrate edge-based (i.e., in a field setting) storage and computing resources capable of characterizing the subsurface, using ambient seismic noise, in near real-time. Our approach enables the continuous assessment of results and the identification of opportunities to modify the sensor array in more optimal configurations depending upon ambient noise field characteristics. Moreover, we can assess the need to leave the array in place for longer or shorter than originally planned with high levels of confidence our survey objectives have been met. Over the course of our four deployments (i.e., Texas – May 2017, Nevada – June 2017, Texas – July 2018, and Nevada – May 2019) we developed an edge-based framework that utilized commercially available communication infrastructure, digitizers, embedded systems, and an established distributed database (i.e., DataStax Enterprise; DSE) to store and process sensor array data, in a field setting. This framework allowed us to overcome real-world performance limiters (e.g., bandwidth, power, etc.) commonly encountered while carrying out remote seismic exploration or monitoring. Moreover, it provides an alternative solution to centralized (i.e., cloud-based) data storage and processing strategies that often have more demanding network capacity and reliability requirements. The use of DSE (powered by Apache Cassandra), as our edge-based distributed database, is central to the scalability and reliability of our framework. To the best of our knowledge, no one has attempted to use DSE or Cassandra, on an embedded system, as a seismic sensor array’s edge-based datastore. Our use of DSE is beneficial in the following three ways: 1) it supports the highly scalable write-heavy workloads common to sensor arrays, 2) it allows for the use of the same fault tolerant distributed database across a variety of commercially available hardware (e.g., embedded systems, servers, etc.), and 3) it seamlessly maintains and replicates data along a user defined continuum of locations (i.e., “edge to cloud”). We believe geoscientists can use our edge-based solution to improve existing and develop novel methods to characterize the subsurface.



Apache Cassandra. DataStax Enterprise. Edge computing. Edge storage. Internet of Things (IoT). Raspberry Pi. Sensor networks. Seismic interferometry.