Computers are becoming increasingly ubiquitous and networked, opening new opportunities for innovative research in the areas of distributed computing and networking. As the number of distributed systems and their range of applicability grows, these systems must act in a more coordinated, autonomous, safe, robust, and efficient manner. IHMC researchers are advancing distributed computing by focusing on three innovative technologies: Agile Computing, KAoS Policy and Domain Services, and Process Integrated Mechanisms. Together, these technologies address such diverse problems as improving the interaction between computational systems in dynamic environments, managing and disseminating information, increasing reliability, efficiency, and power utilization of computational infrastructure, and providing novel architectures for coordinating teams of robots and humans.
The distributed computing group at IHMC is currently working with all three branches of the U.S. Military – the Army, Navy, and Air Force, which have similar but unique challenges. By woking across all three agencies, we can share research and leverage complimentary efforts as well as facilitate collaboration.
The tactical edge network environment has unreliable connectivity, limited bandwidth, and variable latency. Several tools developed by IHMC are in use to address those challenges, and researchers continue to make improvements. The Mockets Communications Library is used by both the Army and the Air Force research laboratories. DisService, a peer-to-peer information dissemination service, recently was incoporated into the Agile Computing Middleware, and is sponsored by both the Army and the Navy. Applications could include supporting information prediction and information on demand.
We are collaborating with the Air Force on their next generation Joint Battlespace Infosphere ( JBI) platform. In this context, we have been developing a dynamic and secure federation capability in order to facilitate information exchange between different enclaves. Federations need a computer architecture that accounts for security and resource management, among other issues. IHMC researchers are developing such a system based on OWL (Web Ontology Language) by extending the KAoS policy services framework.
Green computing is a relatively new effort this year and has been focused on conserving energy in data centers with novel approaches extending agile computing. We have begun establishing a 100-node testbed that will support reasonable-scale experimentation with green computing as well as many of our other technologies.
Finally, many scenarios envision large teams of semiautonomous robots engaged in complex tasks requiring coordinated action to achieve complex and changing goals in an uncertain and possibly hostile environment. Such scenarios could include search and rescue, space exploration, or persistent surveillance. The dynamic patterns of cooperation and action between team activities require coherent planning of team activities. At IHMC, researchers are developing a novel architecture, the Process Integrated Mechanism (PIM), which has a single coordinating authority but avoids the structural difficulties that have traditionally led to its rejection in such complex settings. PIM improves on previous models with regard to coordination, security, predictability, ease of software development, and robustness.