Analysis Performance Characterics of Dynamic Signature File Methods

  • Yoo, Jae-Soo (Dept. of Computer and Communication, Chungbuk National University) ;
  • Choi, Kil-Seong (Dept. of Computer and Communication, Chungbuk National University) ;
  • Kim, Myoung-Ho (Dept. of Computer Science, Korea Advance Institute of Science Technology)
  • Published : 1997.08.01

Abstract

With rapid increase of information requirements from various application areas, there has been much research on dynamic information storage structures that effectively support insertions, deletions and updates. In this appear we evaluate the performance of he existing dynamic signature file methods such as the S-tree, Quick Filter and HS file and provide guidelines for the most effective usage to a given operational environment. We derive analytic performance evaluation models of the storage structures based on retrieval time, storage overhead and insertion time. We also perform extensive experiments with various data distributions such as uniform, normal and exponential distributions. The relationships among various performance parameters are throughly investigated. We show through performance comparison based on analytic models and experiments that regardless of data distribution, the HS file significantly improves performance in both the retreival time and the storage overhead over S-tree and Quick Filter.

Keywords

References

  1. ACM SIGIR Multikey Access Methods Based on Term Discrimination and signature Clustering J.W.Chang;J.H.Lee;Y.J.Lee
  2. IEEE Trans. on Software Engineering v.SE-10 no.2 Design Considerations for a Message File Server S.Christodoulakis;C.Faloutsos
  3. VLDB '84 Development of a Multimedia information system for an Office Environment S.Christodoulakis(et al.)
  4. ACM SIGMOD Issues in the Architecture of a Document Archiver using Optical Disk Technology S.Christodoulakis
  5. ACM SIGIR S-tree: A Dynamic Balanced Signature Index for Office Retrieval U.Deppisch
  6. Proc. of the Fall Joint Computer Conf. An Information Retrieval System Based on Superimposed Coding J.R.Files;H.D.Huskey
  7. Computing Surveys v.17 no.1 Access methods for Text C.Faloutsos
  8. ACM Trans. on Information Systems v.5 no.3 Description and Performance File Method for Office Filing C.Faloutsos;S.Christodoulakis
  9. IEEE Computer Society Technical Committee on Data Engineering v.13 no.1 Signature-based Text Retrieval Methods : A Survey C.Faloutsos
  10. IEEE Trans. on Knowledge and Data Engineering v.4 no.3 Frame-Sliced Signature Files Z.Lin;C.Faloutsos
  11. Proc. IEEE 67 Partial Match Retrieval via the Method of the Superimposed Codes C.S.Roberts
  12. ACM Trans. on Database Systems v.12 no.4 Multikey Access Methods based on Superimposed Coding Techniques R.Sacks-Davis;K.Ramamohanarao
  13. 5th Annual European Computer Conference Selecting Signature Files for Specific Applications P.Tiberio;P.Zezula
  14. IEICE Trans. on Information and Systems v.E76-D no.2 Performance Evaluation of Signature-Based Access Mechanisms for Efficient Information Retrieval J.S.Yoo;J.W.Chang;Y-J.Lee;M.H.Kim
  15. 5th International Conference, DEXA '94 The HS File: A New Dynamic Signature File Method for Efficient Information Retrieval J.S.Yoo;Y.J Lee;J.W.Chang;M.H.Kim
  16. ACM Trans. on Information Systems v.9 no.4 Dynamic Partitioning of Signature Files P.Zezula;F.Rabitti;P.Tiberio