Design of new scheduling algorithm LLF_DM and its comparison with existing EDF, LLF, and DM algorithms for periodic tasks

The most challenging part of scheduling in real time systems is to achieve successful completion of a job before its deadline. Mainly two categories of algorithms i.e. static and dynamic tried to achieve this but both categories failed either in under-loaded condition or in over-loaded condition. Dynamic algorithms achieve optimum results in under-loaded condition but fail to achieve the same in over-loaded condition. On the other side static algorithms do not achieve optimum performance in underloaded condition but perform well in over-loaded condition. So our idea behind designing new scheduling algorithm is to achieve optimum performance in under-loaded condition and to achieve high performance in over-loaded condition. To achieve this we schedule jobs according to dynamic scheduling algorithm LLF (Least Laxity First) when system is under-loaded and when system becomes overloaded we schedule jobs according to static algorithm DM (Deadline Monotonic). In this paper we have proposed a LLF_DM algorithm which achieves optimum performance in under-loaded condition and achieves very high performance in over loaded condition.