Timing and Synchronization for Network Centric Systems

In this Phase I project, we present an approach for high performance synchronization with fault tolerant capabilities suitable for large-scale network centric system architectures. Our approach addresses the significant challenges faced by a time service in a distributed system, particularly for heterogeneous assets. The technique creates a foundation by supplying methods for time synchronization into coherent groups (clusters). The cluster-based synchronization technique adapts well to the multi-level network scheme of a network-centric battlespace system. The cluster model is shown to a viable means of demonstration the utility, efficiency, and scalability of the approach. This approach leads to more efficient results than fully connected structures for (a) graph complexity reductions, (b) reduction in algorithmic overheads and (c) tightness of synchronization. The overall system of clusters can be synchronized to extremely tight margins (<10ns) that can be customized by adjusting the rate of synchronization messages between and across clusters.