SmartGrids: MapReduce Framework using Hadoop

Smart Grids (SGs) are developing as an encouraging technology implied to confront with the energy efficiency issue, presently supported in traditional electrical grids, by disseminating important information in a real-time mode among the various SG unit. The Hadoop framework has been advanced to effective growth of comprehensive data in MapReduce applications. Hadoop users define the application calculation logic in terms of a mapping and a reduction work, which are often described as MapReduce applications. The big data analytics association has authorized MapReduce as a programming model for transforming extensive data on distributed systems. In the Hadoop distributed file systems (HDFS), the MapReduce application data is stored on the Hadoop cluster nodes called DataNodes, and NameNodes control all Datanodes. The audit log files that generates from Advanced metering infrastructure (AMI) in Smart grids would bring about the generation of large bulk of data, i.e. Big Data. In Smart grids, the log data is repeatedly generated as a stream of received and sent packet data. In this paper, we presented Hadoop-MapReduce framework where the audit log files (Big Data) are stored in a Hadoop environment using Map-Reduce technique. The Smart grid under surveillance generates Gigabytes of data (log files) which becomes an issue of storage limitation. This data are mapped and reduced into few Kilobytes or Megabytes. Hence, this technique enables Big Data to store very efficiently. The MapReduce algorithm is executed and our experimental results show significant improvement based on our presented Hadoop-MapReduce framework.