Reviewing the security surveillance of AMI using big data analytics

Abstract

Advanced Metering Infrastructure (AMI) is a kind of communication infrastructure with millions of Smart Meters. The Smart Meters and other components of AMI generate data with high capacity and rate. In the result, data becomes hard to analyze with traditional methods, therefore, some advanced analytics like big data analytics can be very expedient here. There are two types of data passed by every communication system, they are actual and network data. Due to enormous size of AMI network, it produces both actual and network data in terabytes or even more. The actual data is collected from AMI at the AMI repository which can be applied for billing, energy forecasting and demand response applications. The network data controls the passage of actual data and can be a good source to examine the security of AMI system. The authors in the paper review the advanced analytics of the network data for detecting the anomalies in the AMI network. The AMI comprises of a firewall at the entrance of the data center which monitors ins and outs of the data based on security rules. In order to increase the efficiency of the firewall, it is proposed to use the big data analytics for advanced surveillance. There are many tools available for big data analytics. Among those, the apache spark is getting popularity because of its fast in memory cluster computing. It features processing of both batch and streamed data. The inclusion of apache spark as the surveillance tool will make the firewall stream processing more efficient. We also propose the use of machine learning algorithms by AMI firewall for better prediction of anomalies. The machine learning libraries are also well supported by apache spark.