Rowe, Neil C.

Today's increasingly effective methods of data collection are creating large amounts of data that cannot be processed and combined with traditional techniques such as integrated databases. We will investigate the options that the Navy has for such “big data” processing problems on data obtained from different modalities or organizations. In this limited-scope project we will use the example of network-monitoring data to find suspicious anomalies that could represent adversary activities. Such big-data problems can be addressed by using cloud services, explicit distributed processing, or supercomputers. The choice depends on such factors as desired data processing rates, acceptable latencies, error rates in the data, and availability of methods for resolving data conflicts. We will provide guidelines for selecting from the options, and make recommendations for current Navy network monitoring.
NPS Naval Research Program
NPS Naval Research Program
Navy
2017