Enhancement in Performance of LLF Real-Time Scheduling Algorithm

Dynamic scheduling has been remained as a challenging problem for real-time operating systems. EDF (Earliest Deadline First) and LLF (Least Laxity First) are dynamic algorithms and has been proved to be optimal algorithm for single processor system when system is preemptive and under loaded. But limitation of any dynamic algorithm is that its performance decreases drastically when system becomes slightly overloaded. On the other side RM (Rate Monotonic) and DM (Deadline Monotonic) are static algorithms and they do not perform optimal but perform well in both conditions (i.e. under loaded & overloaded). The objective of our dissertation work is to get optimum performance in under loaded condition and to improve the performance of real-time systems in overloaded conditions as well.