Plug-and-play Analytical Framework for Distributed Structured and Unstructured Data Sets for Condition Based Maintenance Plus (CBM+)

Prior to the advent of Big Data and the availability of affordable digital data storage, organizations would typically gather only the data from system logs that could be immediately identified as useful. With modern Big Data Collection, the focus has shifted to collecting and storing as much data as possible; even if that data is not immediately identified as useful to the human eye. The intention is to allow predictive algorithms determine which data is useful and to have a massive stash of data in its native format in digital storage forever, ultimately producing end product reports or notifications to the end user. Our true innovation for this effort will be introducing the hadoop cluster framework as the main data center application vice traditional data center approaches that are human-deterministic. The development approach is to oversimplify Plug and Play functionality through: reuse of native interfaces, cost effective scalability, and preservation of stored data via replicated nodes of the hadoop cluster. The resulting automated data repository will enable new non-traditional predictive algorithms to be developed and its results presented to shipboard users while underway and shore site users for specific analysis as required.