Browse Prior Art Database

Autonomic Install Integrity Verification and Recovery using Parity Files

IP.com Disclosure Number: IPCOM000034885D
Original Publication Date: 2005-Jan-19
Included in the Prior Art Database: 2005-Jan-19
Document File: 2 page(s) / 46K

Publishing Venue

IBM

Abstract

Disclosed is a software component (which can be re-used for any application with some customization and configuration) that can be utilized to verify and repair the application?s files in the event of a failed startup attempt. This is accomplished through the compilation of checksum and parity information either by the vendor or by the user, which is used in the event of an application initialization error to verify and repair any damaged application files.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 52% of the total text.

Page 1 of 2

Autonomic Install Integrity Verification and Recovery using Parity Files

Disclosed is a software component (which can be re-used for any application with some customization and configuration) that can be utilized to verify and repair the application' s files in the event of a failed startup attempt. Since software is largely comprised of static code libraries, graphics, and documentation, a large majority of an installation can be protected with a vendor-generated component configuration that works out of the box. In addition, this configuration can be updated after installation and configuration to include files created or changed for this particular installation. The means by which the verification and recovery is done is through the use of parity information, similar to that of Error Correction Code (ECC) algorithms.

    The typical solution to this sort of problem is redundancy, both in the function and the application data. By replicating the function or service with a cluster of machines, the problem can be circumvented. However, due to the fact that this circumvention doesn't address the root issue, which is corrupted data, a cluster is merely a workaround. The disclosed solution isn't comparable with the clustered approach because it actually solves the problem that the clustered approach merely circumvents. In fact, this component could be used in conjunction with the redundancy of a cluster as an autonomic healing step for a cluster member that has gone down while the other cluster members keep the service available. Rather than simply avoiding the problem and waiting for human intervention, this solution can bring things back to perfect working order without any human intervention.

    This solution is superior to the redundancy of RAID because it can recover from the common class of errors where loss of the file is due to a command given to the system rather than chance corruption.

    This solution is also superior to the common sense approach of keeping a second copy of the application files available to the verification process in order to copy them into place when missing or corrupted. By utilizing the concept of a parity file we gain more fault tolerance, the ability to utilize less storage resources, as well as a more common and maintainable framework.

    When a software application wishing to utilize this functionality is packaged, a list of files and there associated checksums are compiled and included with the installation. In addition, a suitable set...