A New Approach to Querying and Processing Data from CERN Accelerator Logging Service

During the previous 10 years, the CERN Accelerator Logging Service has evolved multiple times i.e. from the expected 1TB of data per year in the beginning to more than 50TB/year. It is used to store and retrieve billions of data acquisitions per day, from across the complete CERN accelerator complex, related subsystems, and experiments. This report includes a description of possible ways of improving the speed of the data retrieval from the CALS service. A short overview of the the possible technologies that can be used i.e. Apache Flume and Apache Spark is given, describing the possible ways in which these two technologies can be implemented.