Distributed Automated Network Clustering framework for Efficient Routing (DANCER)

Navy Warfighters need to quickly communicate situational awareness and command and control (C2) data across large and complex networks to achieve mission goals. The current approach of mass delivery of all data to each participating network node cannot accommodate the throughput the mission requires as the network size increases. To ensure that network resources are allocated efficiently, the Navy needs a solution that automatically clusters network nodes participating in specific information flows based on their geographic locations and the data types relevant to their mission roles. To address these requirements, Charles River Analytics is pleased to propose a Distributed Automated Network Clustering framework for Efficient Routing (DANCER). Under DANCER, we will: (1) design an automatic cluster generator that assigns nodes to clusters such that all end-to-end mission-specific quality of service requirements are satisfied; (2) design a decentralized control data store that provides all nodes resilient access to up-to-date clustering assignments and data needed to dynamically adjust clusters as mission requirements, network structure, and communication patterns change; (3) design a distributed network state monitor that continuously collects observations of network state to provide dynamic input to the clustering algorithm; and (4) provide a feasibility analysis of the DANCER approach.