Value-driven resource assignment in object-oriented real-time dependable systems

Abstract

Abstract

A generalized analytic performance model of distributed systems that perform N tasks using p fault-p

“…Then the performance of a system is determined by the total accumulated reward for a specified time period. Thus one can capture the effective performance of a system, its minimum reliability [17] and optimal utilization of available system resources [3]. The use of performability metrics to evaluate a realtime fault-tolerant system through determining the redundancy levels of tasks is presented in e.g.…”

The meaning and role of value in scheduling flexible real-time systems

“…The work of Bondavalli et al [5] uses value to statically determine the services that should execute in any particular mode to cope with run-time overload conditions. The system is assumed to execute in a reasonably limited number of modes and the attributes characterising the services are constant in any particular mode: (a) if the service execution violates its time constraints (including the case of non-execution at all), a value l i is associated to this event; (b) if a correct (both in time and value) outcome is produced, a value g i is associated to this event; (c) if a timely but erroneous output is produced, a value p i is associated to this event.…”

Realize: resource management for soft real-time distributed systems

“…Bondavalli, Giandomenico and Mura [2] have defined a planner component based on a valuedriven admission policy that minimizes the effects that high reliability requirements have on the expected quality of service and achieves the optimal utilization of available system resources.…”