Hardware Reference
In-Depth Information
9
HANDLING OVERLOAD
CONDITIONS
9.1
INTRODUCTION
This chapter deals with the problem of scheduling real-time tasks in overload condi-
tions; that is, in those critical situations in which the computational demand requested
by the task set exceeds the processor capacity, and hence not all tasks can complete
within their deadlines.
Overload conditions can occur for different causes, including the following:
Bad system design. If a system is not designed or analyzed under pessimistic
assumptions and worst-case load scenarios, it may work for most typical situ-
ations, but it can collapse in particular peak-load conditions, where too much
computation is requested for the available computational resources.
Simultaneous arrival of events. Even if the system is properly designed, the si-
multaneous arrival of several “unexpected” events could increase the load over
the tolerated threshold.
Malfunctioning of input devices. Sometimes hardware defects in the acquisition
boards or in some sensors could generate anomalous sequences of interrupts,
saturating the processor bandwidth or delaying the application tasks after their
deadlines.
Unpredicted variations of the environmental conditions could generate a compu-
tational demand higher than that manageable by the processor under the specified
timing requirements.
Operating system exceptions. In some cases, anomalous combination of data
could raise exceptions in the kernel, triggering the execution of high-priority han-
dling routines that would delay the execution of application tasks.
Search WWH ::
Custom Search