Efficient on-line schedulability test for feedback scheduling of soft real-time tasks under fixed-priority

RODRIGO SANTOS; GIUSEPPE LIPARI; ENRICO BINI

Saint Louis (MO) EEUU

Conferencia; 14th IEEE Real-Time and Embedded Technology and Applications Symposium; 2008

IEEE

When dealing with soft real-time tasks with highly variable  execution times in open systems, an approach that is becoming  popular is to use feedback scheduling techniques to dynamically  adapt the bandwidth reserved to each task.  According to this model,  each task is assigned an adaptive reservation, with a variable  budget and a constant period. The response times of the jobs of the  task are monitored and if different from expected (i.e.  much larger  or much shorter than the task relative deadline), a feedback control  law adjusts the reservation budget accordingly.  However, when the  feedback law algorithm demands an increase of the reservation  budget, the system must run a schedulability test to check if there  is enough spare bandwidth to accommodate such increase.  The  schedulability test must be very efficient, as it may be performed  at each budget update, i.e. potentially at each instance of a task.  In this paper, we tackle the problem of performing an efficient  on-line schedulability test for Resource Reservation systems  implemented through the Sporadic Server on Fixed Priority  scheduling. We propose five different tests with different  complexity and performance. In particular, we propose a novel  on-line test, called emph{Spare Pot algorithm} which shows a good  cost/performance ratio.