Federated Edge Computing for Disaster Management in Remote Smart Oil Fields

Oil & Gas (O&G) industry is extending the extraction operation to remote offshore sites. Cost-effective, efficient, and nature-friendly oil extraction is a challenging issue in these remote sites, due to the disaster-prone nature of oil extraction process and hurdles in accessing these sites. To overcome these difficulties, smart oil fields use numerous sensors (e.g., pipeline pressure, gas leakage, temperature sensors) and can generate more than a terabyte of data per day. The data are transferred to cloud datacenters via high-latency and unstable satellite communication, which is not suitable for latency-intolerant (urgent) disaster-related tasks. Edge computing can be deployed in oil rigs to process the latency-intolerant tasks, however, processing capacity of an edge system falls short at the time of a disaster-when several coordinated activities must be processed within a short time. To address this shortage, we propose robust smart oil fields operating based on a federation of edge computing systems, provisioned from nearby/mobile micro datacenters. Our solution achieves robustness by capturing uncertainties exist both in communication and computation of the federated environment and allocating urgent tasks so that the likelihood of their on-time completion is maximized. Evaluation results reflect significant performance improvement (up to 27%) of the proposed solution when compared to conventional solutions for smart oil fields.