MINT: Mathematical Reliability Models for Energy-Efficient Parallel Disk Systems
Final Report
Our Mission:
The MINT project aims at developing mathematical reliability models for fault-tolerant energy-aware disk systems. Reliability models, which are used to estimate reliability, have been important tools in the design and development of fault-tolerant computer systems. In the past decade, a variety of practical and useful reliability models have been constructed for disk systems. However, most of these models were developed for non-energy efficient disk systems, thereby making it difficult to apply the existing reliability models to energy-aware disk systems. Therefore, the overall objective of this project is to address the mathematical underpinnings of modeling reliability of energy-efficient parallel disk systems, where fault tolerance and energy-saving techniques will be seamlessly integrated together to conserve energy without sacrificing reliability in parallel disk systems.
Our Approach:
The project will contribute to reliability modeling techniques for fault-tolerant and energy-efficient parallel disk systems by developing a reliability analysis modeling toolkit accompanied with a set of novel mathematical reliability models. The innovative models investigated in this project include disk power consumption models, a reliability model for parallel disk systems with redundancy techniques, a reliability model for repairable and energy-efficient parallel disk systems, a fault recovery model for energy-efficient parallel disk systems. The proposed research has three main strengths. First, we bridge the technology gap between reliability analysis and energy conservation techniques in the context of parallel disk systems. Second, we seek to develop an array of mathematical reliability models for energy-efficient parallel disk systems. Finally, we will implement a reliability analysis toolkit, which will be the first toolkit of its kind designed specifically to study a variety of fault-tolerant and energy-saving techniques.
People:
Faculty
Xiao Qin (PI), Assistant Professor, Auburn University
Intel Technical Staff
Yiming Yang, Senior Software Engineer, Intel Corporation
Students
Kiranmai Bellam [Ph.D. Student | Awards Total: 3 | Publications Total: 15; Journal Papers: 5; Conference Papers: 10]
Dissertation Research: Fault Tolerance, Energy, and Security Issues in Real-Time Systems.
Fault tolerance, security, and energy issues in modern real-time systems are of critical importance. This work is intended to seamlessly integrate security services and energy conservation techniques for real-time systems while endeavoring to achieving high system reliability. Read more...
Adam Manzanares [Ph.D. Student | Awards: 2 | Publications Conference Papers: 1]
Dissertation Research: Energy Efficient Wireless Storage Systems.
This study addresses the issue of modeling energy-efficient wireless storage systems and wireless network interconnections. Adam's work on energy conservation for high-performance clusters appeared in the proceedings of the 8th IEEE International Conference on Cluster Computing (Cluster'06), Sept. 2006. Read more...
Fangyang Shen [07/2007 - Present | Publications Total: 3; Journal Papers: 0; Conference Papers: 3]
Ph.D. Candidate, Auburn University.
Research Project: Reliability Models of Energy-Efficient Parallel Disk Systems with Data Mirroring.
In this project we make use of a Markov process to develop a quantitative reliability model for energy-efficient parallel disk systems with data mirroring. With the new models in place, a reliability analysis tool is developed to efficiently evaluate reliability of fault-tolerant parallel disk systems with multiple power states.
Our Funding:
MINT is being funded through the Computer Systems Research (CSR) award CNS-0757778 from the U.S. National Science Foundation (NSF).
Contact Us
Xiao Qin, Ph.D., Principle Investigator
Shelby Center for Engineering Technology, Suite 3101E
Department of Computer Science and Software Engineering
Samuel Ginn College of Engineering
Auburn University , AL 36849-5347
Office: 1-334-844-6327
Fax: 1-334-844-6329
Email: xqin@auburn.edu