INFORMATION SCIENCES 59, 165-187 (1992) A General Approach to the Identification of Compensated Faults in Robust Data Structures* A. RAVICHANDRAN and K. KANT zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA Department of Computer Science, Pennsylvania State University, University Park, Pennsylvania 16802 165 ABSTRACT In this paper we give an optimal algorithm for the identification of faulty attributes in a robust data structure. The algorithm does not use any fault syndrome table since the size of such a table could be large, particularly when faults can compensate one another arbitrarily. The data structure is viewed as a collection of data elements related via some “attributes.” The relationships are specified by a set of axioms in first order logic. Faults in attributes invalidate some of the axioms. The invalidated axioms are used to identify the faulty attributes. We show that the identification is possible in time proportional lo the number of axioms even when faults compensate one another arbitrarily. This is optimal, zyxwvutsrqponmlk since our method of axiom generation does not yield any redundant axioms. 1. INTRODUCTION Redundancy is essential for detecting, correcting, and recovering from hardware and/or software errors. Redundancy can be in the form of backup copies or embedded in the representation of data itself. A robust data structure employs redundancy to enable correction in a faulty instance. Identi- fication of faulty attributes of the data structure is a prerequisite for correc- tion of these faulty attributes. There has been a great deal of effort focused on designing robust data structures and the correction algorithms for these robust data structures [2-61. The correction algorithms given in [2-61 are designed for *This research was supported by Applied Research Lab at Pennsylvania State University under Grant No. E/F 1915. DElsevier Science Publishing Co., Inc. 1992 655 Avenue of the Americas, New York, NY 10010 0020-0255/92/$03.50