Task Allocation in heterogeneous Distributed Real Time System for Optimal Utilization of Processor’s Capacity

Abstract

In Distributed Real Time System (DRTS), systematic allocation of the tasks among processors is one of the important parameter, which determine the optimal utilization of available resources. If this step is not performed properly, an increase in the number of processing nodes results in decreasing the total system throughput. The Inter-Task Communication (ITC) is always the most costly and the least reliable parameter in the loosely coupled DRTS. In this paper an efficient task allocation algorithm is discussed, which performs a static allocation of a set of "m" tasks T = {t 1 ,t 2 ,…t m } of a program to a set of "n" processors P = {p 1 ,p 2 ,….p n }, (where, m >> n) to minimize the application program's Parallel Processing Cost(PPC) with the goal to maximize the overall throughput of the system through and allocated load on all the processors should be approximately balanced. While designing the algorithm the Execution Cost (EC) and Inter Task Communication Cost (ITCC) have been taken into consideration.

Hybrid model for tasks scheduling in distributed real time system

“…Approaches Environment Attiya and Hamam (2006) Discussed a simulating annealing algorithm for raising system reliability Heterogeneous Akbari and Rashidi (2016) To manage tasks scheduling, a technique is presented to diminish execution time whilst allowing for maximal parallelization Heterogeneous Anderssont et al (2001) Proposed a scheduling algorithm, simple extension of the uniprocessor ratemonotonic scheduling algorithm, for static-priority scheduling Heterogeneous Elsadek and Wells (1999) Gives the idea of a HMLM/SA heuristic model, which performs allocation of modules in a style that is arranged to diminish parallel execution time of program Heterogeneous Kopiddakis et al (1997) Two new heuristic algorithms to minimize inter-task communication time and processors execution have been presented Heterogeneous Lo (1988) Heuristic algorithms have been given for scheduling of tasks in distributed system Heterogeneous Lee et al (1992) Stone's network flow technique is extended to the situation for a linear array of any number of processors Heterogeneous Shatz et al (1992) Introduced an allocation idea for tasks for raising reliability of system Heterogeneous Singh and Garg (2014) An efficient scheduling technique is discussed with objective to maximize complete system throughput and assigned load onto all processors is approximately balanced Heterogeneous Topcuoglu et al (2002) Investigated two scheduling methods for a fixed number of processors with two objectives as quick scheduling time and to obtain high performance, which are called CPOP and HEFT algorithms Heterogeneous Ucar et al (2006) Considered an issue in which tasks have to be allocated to processors with identical links to diminish the sum of communication and total execution times Heterogeneous Yadav et al (2011a) The ANN based model is discussed that provides optimal solution and assigned load is balanced corresponding to average load onto all processors Heterogeneous attributes into known number of group. Actually, clustering is applied to diminish a set of huge data into a relatively compact analysis.…”

Task Allocation Model Based on Hierarchical Clustering and Impact of Different Distance Measures on the Performance

“…Ma and Lee (1982) presented a model of taskallocationdependonazero-oneprogrammingapproachandthebranchandboundmethodin distributedcomputing. SarjeandSagar(1991) Singh and Garg (2014) derived a systematic allocation model of tasks to diminish the program's Parallel Processing Cost (PPC) with the hope to raise the overall system's throughput. Several approaches to the dynamic tasks allocations have been analyzedinthebygonetime.…”