A Simple MPI Process Swapping Architecture for Iterative Applications

Sievert, Otto; Casanova, Henri

doi:10.1177/1094342004047430

Cited by 30 publications

(15 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…SALSA actors do not leave any residual dependencies [7] or create resource conflicts when moving around computing nodes. This approach is effective because it is light-weight compared to other migration techniques such as VM-level migration [8] and is able to keep the workers running without restarting the application as opposed to check-point restart approaches [9].…”

Section: B Workload Scalability Over a Hybrid Cloudmentioning

confidence: 99%

Accurate Resource Prediction for Hybrid IaaS Clouds Using Workload-Tailored Elastic Compute Units

Imai

Chestna

Varela

2013

2013 IEEE/ACM 6th International Conference on Utility and Cloud Computing

View full text Add to dashboard Cite

Abstract-Cloud computing's pay-per-use model greatly reduces upfront cost and also enables on-demand scalability as service demand grows or shrinks. Hybrid clouds are an attractive option in terms of cost benefit; however, without proper elastic resource management, computational resources could be overprovisioned or under-provisioned, resulting in wasting money or failing to satisfy service demand. In this paper, to accomplish accurate performance prediction and cost-optimal resource management for hybrid clouds, we introduce Workload-tailored Elastic Compute Units (WECU) as a measure of computing resources analogous to Amazon EC2's ECUs, but customized for a specific workload. We present a dynamic programming-based scheduling algorithm to select a combination of private and public resources which satisfy a desired throughput. Using a loosely-coupled benchmark, we confirmed WECUs have 24% better runtime prediction ability than ECUs on average. Moreover, simulation results with a real workload distribution of web service requests show that our WECU-based algorithm reduces costs by 8-31% compared to a fixed provisioning approach.

show abstract

Section: B Workload Scalability Over a Hybrid Cloudmentioning

confidence: 99%

Accurate Resource Prediction for Hybrid IaaS Clouds Using Workload-Tailored Elastic Compute Units

Imai

Chestna

Varela

2013

2013 IEEE/ACM 6th International Conference on Utility and Cloud Computing

View full text Add to dashboard Cite

show abstract

“…Moreover, restarting the application can incur expensive startup costs, and significant application modifications may be required for specialized restart code. Our process swapping approach, which was initially described in, (24) provides an alternative that is lightweight and easy to use, but less flexible than our migration approach.…”

Section: Rescheduling By Processor Swappingmentioning

confidence: 99%

New Grid Scheduling and Rescheduling Methods in the GrADS Project

Berman

Casanova

Chien

et al. 2005

Int J Parallel Prog

127

View full text Add to dashboard Cite

“…These applications were integrated into the GrADS framework and tested in various experiments as part of the following efforts: ScaLAPACK [16], Jacobi [17], Game of Life [17], Fish [18], and FASTA [19].…”

Section: Application Running On Emulated Environmentmentioning

confidence: 99%

Validating and Scaling the MicroGrid: A Scientific Instrument for Grid Dynamics

2004

View full text Add to dashboard Cite

Large-scale Grids that aggregate and share resources over wide-area networks present major challenges in understanding dynamic application and resource behavior for performance, stability, and reliability. Accurate study of the dynamic behavior of applications, middleware, resources, and networks depends on coordinated and accurate modeling of all four of these elements simultaneously.We have designed and implemented a tool called the MicroGrid which enables accurate and comprehensive study of the dynamic interaction of applications, middleware, resource, and networks. The MicroGrid creates a virtual Grid environment -accurately modeling networks, resources, the information services (resource and network metadata) transparently. Thus, the MicroGrid enables users, Grid researchers, or Grid operators to study arbitrary collections of resources and networks. The MicroGrid includes the MaSSF online network simulator which provides packet-level accurate, but scalable network modeling.We present experimental results with applications which validate the implementation of the MicroGrid, showing that it not only runs real Grid applications and middleware, but that it accurately models both their and underlying resource and network behavior. We also study a range of techniques for scaling a critical part of the online network simulator to the simulation of large networks. These techniques employ a sophisticated graph partitioner, and a range of edge and node weighting schemes exploiting a range of static network and dynamic application information. The best of these, profile-driven placement, scales well to online simulation of large networks of 6,000 nodes using 24 simulation engine nodes.

show abstract

A Simple MPI Process Swapping Architecture for Iterative Applications

Cited by 30 publications

References 9 publications

Accurate Resource Prediction for Hybrid IaaS Clouds Using Workload-Tailored Elastic Compute Units

Accurate Resource Prediction for Hybrid IaaS Clouds Using Workload-Tailored Elastic Compute Units

New Grid Scheduling and Rescheduling Methods in the GrADS Project

Validating and Scaling the MicroGrid: A Scientific Instrument for Grid Dynamics

Contact Info

Product

Resources

About