Document Type

Thesis - University Access Only

Award Date

2007

Degree Name

Master of Science (MS)

Department / School

Electrical Engineering and Computer Science

Abstract

The checkpoint and recovery facility saves the state of process to a stable storage periodically and recovers the user program by using the latest checkpoint when a process has failed. Since the checkpoint and recovery facility prevents the loss of intermediate results generated by long-running processes due to failure, it is very important for long running processes [l]. Various versions of the checkpoint and recovery facility [2, 3] were proposed to improve the performance of the checkpoint and recovery facility, but the checkpoint and recovery facility which is compatible with Linux kernel 2.6 is not implemented yet. The necessity of the checkpoint and recovery facility in Linux kernel 2.6 is gradually increasing, because Linux kernel 2.6 is currently used as a purpose of development in various fields. In this thesis, I develop the Nckpt, a new version of the checkpoint and recovery facility, on Linux kernel 2.6. Nckpt provides a user-directed checkpoint and compatibility with Linux kernel 2.6. The main concept of the Nckpt mainly comes from the old-version of the checkpoint and recovery facility. Experimental evaluation shows that my implementation is quite efficient and works correctly.

Library of Congress Subject Headings

Linux

Fault-tolerant computing

Software engineering

Format

application/pdf

Number of Pages

66

Publisher

South Dakota State University

Share

COinS