VPSF: A Parallel Signature File Technique Using Vertical Partitioning and Extendable Hashing

Abstract

In this paper, we propose a Vertically-partitioned Parallel Signature File (VPSF) method which can partition a signature file vertically. Our VPSF method uses an extendable hashing technique for dynamic environment and uses a frame-sliced signature file technique for efficient retrieval. Our VPSF method also can eliminate the data skew and the execution skew by allocating each frame to a processing node. To prove the efficiency of our VPSF method, we compare its performance with those of the conventional parallel signature file methods, i.e., HPSF and HF, in terms of retrieval time, storage overhead, and insertion time. The experiment runs on several distributions with normal, half, and dou- ble standard deviations of the real data. The result shows that our VPSF achieves about 40% better retrieval performance than the HF in all cases. In addition, we show that our VPSF gains about 20~50% improvement in retrieval time, compared with the HF and HPSF on record sets with the half deviation. As a result, our VPSF generally outperforms on retrieval performance when the records of a database are uniform in size.