Deadline-constrained workflow scheduling in volunteer computing systems

One of the main challenges in volunteer computing systems is scheduling large-scale applications expressed as scientific workflows. This work aims to integrate partitioning scientific workflows and proximity-aware resource provisioning to increase the percentage of workflows that meet the deadline in peer-to-peer based volunteer computing systems. In the partitioning phase, a scientific workflow is partitioned into sub-workflows in order to minimize data dependencies among them. We utilize knowledge-free load balancing policy and proximity of resources to distribute sub-workflows on volunteer resources. Simulation results show that the proposed workflow scheduling system improves the percentage of scientific workflows that meet the deadline with average of 18% under a moderate workload