A New Task Scheduling Algorithm in Distributed Computing Environments

Distributed computing environments are well suited to meet the computational demands of diverse groups of tasks. At present the most popular model characterizing task’s precedence is DAG(directed acyclic graph). In [2], a novel model called TTIG(Temporal Task In Interaction Graph)that is more realistic and universal than DAG and its corresponding algorithm called MATE are presented. This paper extends TTIG model and proposes a new static scheduling algorithm called GBHA(group-based hybrid algorithm) that eliminates cycles when traversing TTIG, so that global information can be captured. Simulation results show that our algorithm outperforms MATE significantly in homogeneous systems.