Autonomic management of replication in voting-based fault-tolerant data collection systems

We outline the methods for autonomic management of device replication in a voting-based data collection system to deal with situations where the available network bandwidth may fluctuate and/or the device fault parameters may change unpredictably. We model the voting system as a `black-box´ with prescriptive I/O behaviors, wherein a management module exercises a macroscopic control of the voting box with situational inputs: such as application priorities, system resource constraints, and external threat levels.