Distributed snapshots: determining global states of distributed systems
ACM Transactions on Computer Systems (TOCS)
Market-based control: a paradigm for distributed resource allocation
Market-based control: a paradigm for distributed resource allocation
The grid
SETI@home: an experiment in public-resource computing
Communications of the ACM
Managing Checkpoints for Parallel Programs
IPPS '96 Proceedings of the Workshop on Job Scheduling Strategies for Parallel Processing
Result checking in global computing systems
ICS '03 Proceedings of the 17th annual international conference on Supercomputing
IPDPS '05 Proceedings of the 19th IEEE International Parallel and Distributed Processing Symposium (IPDPS'05) - Workshop 1 - Volume 02
Analyzing Market-Based Resource Allocation Strategies for the Computational Grid
International Journal of High Performance Computing Applications
| Hi-index | 0.00 |
Client grid computing models based on participation of non-dedicated clients have been popular for computationally intensive tasks. Two fundamental requirements of these models are efficiency and accuracy. Common implementations use 1) checkpointing mechanisms for higher efficiency and 2) redundancy to achieve accurate results. In this paper, we formulate client grid computation using stochastic models and analyze the effects of checkpointing and redundancy in relation to performance. We first quantify the computation times required for a task with and without checkpointing, then the relationship between result accuracy and redundancy. Finally, we give a sensitivity analysis for parameters relating to client availability, checkpointing, and redundancy to provide guidelines on design and implementation of client grid systems.