Efficient fingerprint lookup using Prefix Indexing Tablet

Abstract

Backups protect the file systems from disk or other hardware failures, software errors that may corrupt the file system and natural disasters. However, a single file may be present as multiple copies in the file system. Hence searching time to find the redundant data and to eliminate them is high. In addition to this, redundant data consumes more space in storage systems. Data de-duplication techniques are used to address these issues. Fingerprint lookup is a key ingredient for efficient de-duplication. This paper proposes an efficient Fingerprint lookup technique called Prefix Indexing Tablets in which the fingerprint lookup is performed only on necessary tablets. Further to reduce the fingerprint lookup delay, only the prefix of the fingerprint is considered. Experimentation on standard datasets show that the lookup latency of the proposed deduplication method is reduced by 62% and the running time is improved.