Faulttolerant Parallel Applications with Dynamic Parallel Schedules.pdfVIP

Fault-Tolerant Parallel Applications with Dynamic Parallel Schedules: A Programmer’s Perspective Sebastian Gerlach, Basile Schaeli, and Roger D. Hersch Ecole Polytechnique F´ed´erale de Lausanne (EPFL), School of Computer and Communication Sciences, Station 14, 1015 Ecublens, Switzerland dps@epfl.ch Abstract. Dynamic Parallel Schedules (DPS) is a flow graph based framework for developing parallel applications on clusters of worksta- tions. The DPS flow graph execution model enables automatic pipelined parallel execution of applications. DPS supports graceful degradation of parallel applications in case of node failures. The fault-tolerance mech- anism relies on a set of backup threads stored in the volatile storage of alternate nodes that are kept up to date by both duplicating transmit- ted data objects and performing periodical checkpointing. The current state of a failed node can be reconstructed on its backup threads by re-executing the application since the last checkpoint. A valid execu- tion order is automatically deduced from the flow graph. The addition of fault-tolerance to a DPS application requires only minor changes to the application’s source code. The present contribution focuses on the development of fault-tolerant parallel applications with DPS from a pro- grammer’s perspective. 1 Introduction and Related Work Clusters of commodity workstations are rapidly growing in size and complexity as computation power requirements increase. The large number of computing nodes incorporated within a cluster dramatically increases the li