Adaptive Monitoring for Analytics Placement in Tactical Networks

For a broad range of C
⁴
ISR applications, the ability to conduct operations in dynamic tactical network environments requires the understanding of resource availability. We propose a framework for efficiently monitoring such a complex network for the purpose of allocation of networking, computing, and analytics resources. The framework maintains awareness of a heterogeneous network, including compute platforms and networking resources, to improve the performance of analytics placement. We outline the framework comprising metric selection, compression, and scheduling. We introduce the concept of network maps, a distributed reporting method where nodes determine their own reporting schedule to maintain analytics placement quality. Next, we present simulation results that show the performance improvement in monitoring and analytics placement. Finally, we describe the monitoring architecture that we are developing to conduct emulation experiments.