Research of Hadoop Task Speculative Execution Based on Process Migration

Yin Cheng Liang; Kai Zhu; Cui Du

doi:10.4028/www.scientific.net/AMM.607.877

Paper Titles

Speeding Maintenance Performance through Time Study
p.860

Ontologies to Capture and Share Knowledge in a Distributed Environment Needed to Support Decision Making in Probabilistic Design
p.864

Dynamic Clustering Heuristic Method for Smart Grid Computing
p.868

Performance Analysis of the PISO-Based CFD Simulation
p.872

Research of Hadoop Task Speculative Execution Based on Process Migration
p.877

Review of Analysis Model on Three-Dimensional Braided Composites
p.881

The Fuzzy Evaluation Based on Entropy Method for Heavy Oil Reservoir Horizontal Well CO₂ Huff and Puff Well Selection
p.886

Numerical Analysis of the Iced Transmission Line Galloping
p.894

Double Erasures Correcting MDS Code with Heterogeneous Placement
p.901

HomeApplied Mechanics and MaterialsApplied Mechanics and Materials Vol. 607Research of Hadoop Task Speculative Execution...

Research of Hadoop Task Speculative Execution Based on Process Migration

Abstract:

Current Hadoop platform task speculative execution can’t guarantee the backup task execution speed. This paper proposes a Hadoop speculative execution mechanism based on process migration. The backup task initializes the Task state with ExecutionState object which is formed according to the original task status. Thus, the backup task can continue the task instead of restarting a new task, this makes sure the backup task is more efficient than the original task. The results of running jobs on the hadoop platform prove that task speculative execution based on progress migration improves the efficiency greatly.