Cloudscape Protection: Securing AMI Meters with Multi-Cloud Strategies in Pakistan

Abstract

AMI is disrupting grid and consumer relationship within the energy sector of Pakistan. But the increased dependence on cloud services also comes with significant security challenges, including data leaks and breaches, unauthorized access, and service outages. To this end, this work advocates for a multi-cloud approach as a powerful solution for security, resilience and performance improvement for AMI systems. By spreading infrastructure across multiple cloud providers, it reduces risk associated with single cloud provider reliance including vendor lock-in and single points of failure. It uses a real data-set on AMI and various machine learning techniques such as Decision Trees, Random Forests, Support Vector Machines to detect anomalies and possible intrusions. Out of these, Decision Trees, with accuracy level as highly as 92.5% and good precision and recall levels, proved the effectiveness of the model for detecting the threat in real-time. The paper also discusses technical, operational and legal challenges faced in implementation of multi-cloud infrastructures in Pakistan which provides perspective on its cost-effectiveness and implementation issues. This study adds to the emerging body of knowledge on smart grid security, as it provides an empirically validated AI-empowered multi clouds security framework for the developing countries. It has the practical value to power utilities, policy makers and researchers who are interested in developing scalable secure smart grid infrastructure.