Enhancing Survivability with Proactive Fault-Containment

Realistic survivable systems must assume that faults will occur within the system. When a malicious fault is activated, it may work to cause damage and to spread; until the system has recovered from this damage, it will have a lower degree of survivability than it did before the fault occurred. By proactively containing faults that would otherwise spread throughout the system, we can reduce the amount of potential damage to the system, and thereby maintain system survivability. Enabling proactive survivability carries with it a number of challenges, including the need to quantify survivability in order to justify the potential overhead of the proactive mechanisms, the need to select appropriate fault detection strategies, and the need to address runtime problems like deciding when and where to focus proactive effort.