Authors:
Damien Hogan
1
;
Tom Arbuckle
1
;
Conor Ryan
1
and
Joe Sullivan
2
Affiliations:
1
University of Limerick, Ireland
;
2
Limerick Institute of Technology, Ireland
Keyword(s):
Genetic Programming, Binary Classifier, Solid State Drive, Flash Memory.
Related
Ontology
Subjects/Areas/Topics:
Artificial Intelligence
;
Computational Intelligence
;
Evolutionary Computing
;
Genetic Algorithms
;
Informatics in Control, Automation and Robotics
;
Intelligent Control Systems and Optimization
;
Knowledge Discovery and Information Retrieval
;
Knowledge-Based Systems
;
Machine Learning
;
Soft Computing
;
Symbolic Systems
Abstract:
Flash memory based Solid State Drives (SSDs) are gaining momentum toward replacing traditional Hard Disk Drives (HDDs) in computers and are now also generating commercial interest from enterprise data storage companies. However, storage locations in Flash memory devices degrade through repeated programming and erasing. As the storage blocks within a Flash device deteriorate through use, their ability to retain data while powered off over long periods also diminishes.
Currently there is no way to predict whether a block will successfully retain data for a specified period of time while powered off. We detail our use of Genetic Programming (GP) to evolve a binary classifier which predicts whether blocks within a Flash memory device will still satisfactorily retain data after prolonged use, saving considerable amounts of testing time. This is the first time a solution to this problem has been proposed and results show an average of over 85% correct classification on previously unseen da
ta.
(More)