Mobile Agent-based Distributed Search and Analytics Tool (MADSAT)

The Army"s intelligence and operational information sources in the era of net-centric warfare and cloud computing are highly distributed and diverse. The challenge is to meet the wide variety of requirements of battle command staff who search for and analyze data for mission- and time-critical decisions. Time criticality, data source autonomy, and bandwidth limitations impose further constraints on transferring relevant data to local computational nodes. Mobile agents take computations at data sources, thus maintaining the autonomy of individual data sources, but security is a prime concern in regards to their free movement. Here we propose to design and develop a Mobile Agent-based Distributed Search and Analytics Tool (MADSAT), where a web interface allows agents to be spawned within a controlled environment for increased security. MADSAT translates a search query to a set of sub-queries via a combination of planning and traditional database query optimization techniques. MADSAT then spawns mobile agents corresponding to sub-queries for retrieving data from appropriate sources and then filters and merges data from these agents and returns the answers. For complex analytics supporting situational awareness and decision-aiding, query decomposition and remote execution of mobile codes are guided by a distributed fusion architecture based on Bayesian Network models.