Fault tolerance via replication in coarse grain data-flow

Nguyen-Tuong, Anh; Grimshaw, Andrew S.; Karpovich, John F.

doi:10.1007/bfb0023066

Cited by 5 publications

(3 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…business analytics. Although each subtask in a parallel job may be highly reliable, the job reliability decreases with the number of tasks, and a single subtask failure will make the whole job fail [1].…”

Section: Introductionmentioning

confidence: 99%

“…This effect is particularly critical for latencysensitive application, for which all subtasks must complete within a strict deadline, in order for the application to feel responsive [5]. Several fault-tolerance approaches [1], [11], [12] have been proposed for processor failures, adopting redundancy in space and time to improve system reliability. Instead, the focus of this paper is on the analysis of systems subject to request failures.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Assessing the Impact of Concurrent Replication with Canceling in Parallel Jobs

Qiu

Pérez

2014

2014 IEEE 22nd International Symposium on Modelling, Analysis &Amp; Simulation of Computer and Telecommunication Systems

View full text Add to dashboard Cite

Parallel job processing has become a key feature of many software applications, e.g., in scientific computing. Parallelization allows these applications to exploit large resource pools, such as cloud or grid data centers. However, a job composed of a large number of parallel tasks will suffer a failure if any of its tasks fail, requiring reprocessing and additional delays. In this paper, we explore the effect that the replication of parallel jobs has on the job reliability and response time, as well as on resource utilization. The replication mechanism consists of concurrently processing replicas, at either the job or the task level, retrieving the results of the replica that finishes first, if any, and canceling any remaining replica in process. We propose a stochastic model that explicitly considers parallel job processing, replication at both the job and the task level, and handles general arrival processes. We develop a numericallyefficient algorithm to solve large-scale instances of the model and compute key performance metrics. We observe that the task cancellation mechanism offers an effective way of limiting the increase in resource utilization, allowing the use of replicas that not only increase the job reliability, but have the potential to reduce the response times.

show abstract

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Assessing the Impact of Concurrent Replication with Canceling in Parallel Jobs

Qiu

Pérez

2014

2014 IEEE 22nd International Symposium on Modelling, Analysis &Amp; Simulation of Computer and Telecommunication Systems

View full text Add to dashboard Cite

show abstract

“…The architecture is general, but allows considerable flexibility in responding to failures of computation nodes if a job has appropriate structure. Specifically, we take advantage of two kinds of structure: (a) sparse dependencies among tasks (in the spirit of traditional data flow models [7]), and (b) similarity of computation products of different tasks. The former is encoded in a dependency graph, while the latter motivates the substitution graph.…”

Section: Introductionmentioning

confidence: 99%

Fault oblivious high performance computing with dynamic task replication and substitution

et al. 2011

View full text Add to dashboard Cite

Traditional parallel programming techniques will suffer rapid deterioration of performance scaling with growing platform size, as the work of coping with increasingly frequent failures dominates over useful computation. To address this challenge, we introduce and simulate a novel software architecture that combines a task dependency graph with a substitution graph. The role of the dependency graph is to limit communication and checkpointing and enhance fault tolerance by allowing graph neighbors to exchange data, while the substitution graph promotes fault oblivious computing by allowing a failed task to be substituted onthe-fly by another task, incurring a quantifiable error. We present optimization formulations for trading off substitution errors and other factors such as available system capacity and low-overlap task partitioning among processors, and demonstrate that these can be approximately solved in real time after some simplifications. Simulation studies of our proposed approach indicate that a substitution network adds considerable resilience and simple enhancements can limit the aggregate substitution errors.

show abstract

Robust Parallel Job Scheduling Infrastructure for Service-Oriented Grid Computing Systems

Abawajy

2005

Computational Science and Its Applications – ICCSA 2005

View full text Add to dashboard Cite

Fault tolerance via replication in coarse grain data-flow

Cited by 5 publications

References 15 publications

Assessing the Impact of Concurrent Replication with Canceling in Parallel Jobs

Assessing the Impact of Concurrent Replication with Canceling in Parallel Jobs

Fault oblivious high performance computing with dynamic task replication and substitution

Robust Parallel Job Scheduling Infrastructure for Service-Oriented Grid Computing Systems

Contact Info

Product

Resources

About