Topology Management Analysis and Algorithms (TMAA)
Navy SBIR FY2008.2

Sol No.: Navy SBIR FY2008.2
Topic No.: N08-198
Topic Title: Topology Management Analysis and Algorithms (TMAA)
Proposal No.: N082-198-0117
Firm: Technology Unlimited Group
1885 Sefton Place
San Diego, California 92107-2623
Contact: John Reddan
Phone: (619) 865-5173
Web Site:
Abstract: TUG's proposed solution to the Topology Management for Directional Antenna-based Networks (DA-Nets) applies the concept of clusters to the formation and optimization of network topologies. The decentralized control approach required by military applications of DA-Nets requires the distribution of information across the network such that each node can calculate the topology based on a shared view of the network and its traffic patterns. A cluster associates nodes performing related functions, which is both a natural basis for creating topologies and a method for reducing the computational complexity of managing DA-Net topology. A valuation function is defined which calculates the value of a network configuration based on utilization requirements (e.g., network traffic). The solution utilizes optimization algorithms to optimize inter and intra-cluster connectivity based on the `value' criteria calculated by the valuation function. The end result is a self-optimizing network topology which is constantly seeking to improve its own `value. The work performed in this effort will: - Examine options for implementing the Valuation Function which assess network value - Tailor existing TUG optimization algorithms for DA-Net application - Define the information and message exchange requirements necessary to perform distributed topology management
Benefits: The TUG Cluster-based Topology Optimization Algorithms (CTOA) will support distributed asynchronous optimization of network topology. The notion of distributed asynchronous optimization can be applied to many problems, including the sensor resource optimization that the core algorithms were designed for. Clustering is the cornerstone concept behind the CTOA approach. The use of clustering is a widely accepted concept in computer science and other fields (e.g., economic development), and has been applied to optimization of key resource utilization (e.g., server clusters, database clusters). The rationale for clustering is that improved results can be obtained by organizing complementary capabilities such that the exchange of information can be facilitated. Clustering makes sense in the context of Directional Antenna-based Networks (DA-Nets) because it:  Reduces the number of connections for nodes requiring large data volumes increases maximum bandwidth  Associates network nodes which are directly communicating with each other thereby reducing the average hop count and latency  Utilizes the long range capabilities of directional antennas to communicate directly with destination nodes rather than hop from nearest neighbor to nearest neighbor TUG extends the inherent advantages of the Cluster concept by using algorithms to continuously optimize the placement of nodes within the clusters. TUG uses a Valuation Function to assess the level of information exchange (e.g., network traffic) such that the degree of complementariness can be compared for different solutions. When the algorithms identify an improved DA-Net configuration, a distributed negotiation process is then used to implement the solution. While aspects of the solution are specific to the DA-Net problem, the optimization concepts can be applied to a much large range of problems.