Charles River Analytics (CRA) has partnered with George Washington University (GWU) to develop a self-healing adaptation infrastructure for loss tolerance (SAIL) which could increase the resilience of US Navy ships.
Navy ships use a complex, interdependent network of computational elements, including servers, sensor processors, and control systems. When a computer running mission software is damaged, it is vital that other ship components pick up the slack.
For example, if an enemy missile destroys the system that processes radar data, the radar sensor hardware may remain operational. In this case, a ship equipped with SAIL could self-heal, adapting its network of surviving hardware and software resources to perform radar data analysis, replacing the lost functionality, the company emphasized.
Specifically, when a processing node on the ship goes dark, SAIL will recognize the failure and distributes the lost capability.
Because a single surviving node may not have enough computational power to run another full application, SAIL allows the application modules comprising the lost capability to be executed on separate, coordinating nodes, according to the firm.
“The advantage of SAIL is that it is decentralized, so platform capabilities can remain available in the presence of failures,” said Gerald Fry, Scientist at Charles River Analytics and Principal Investigator of the SAIL effort.
“SAIL also lets analysts track resource allocations, so they know whether to add low-SWaP compute nodes to cover potential outages or remove resources from components that were underutilized.”