Journal of KIISE:Databases (한국정보과학회논문지:데이타베이스)

Korean Institute of Information Scientists and Engineers (한국정보과학회)
권호 표지

KDBcs-Tree : An Efficient Cache Conscious KDB-Tree for Multidimentional Data

KDBcs-트리 : 캐시를 고려한 효율적인 KDB-트리

  • 여명호 (충북대학교 정보통신공학과) ;
  • 민영수 (한국전자통신연구원 홈네트워크연구단) ;
  • 유재수 (충북대학교 전기전자컴퓨터공학부)
  • Published : 2007.08.15
Download PDF
⟨ Previous Next ⟩

Abstract

We propose a new cache conscious indexing structure for processing frequently updated data efficiently. Our proposed index structure is based on a KDB-Tree, one of the representative index structures based on space partitioning techniques. In this paper, we propose a data compression technique and a pointer elimination technique to increase the utilization of a cache line. To show our proposed index structure's superiority, we compare our index structure with variants of the CR-tree(e.g. the FF CR-tree and the SE CR-tree) in a variety of environments. As a result, our experimental results show that the proposed index structure achieves about 85%, 97%, and 86% performance improvements over the existing index structures in terms of insertion, update and cache-utilization, respectively.

본 논문에서는 데이타의 갱신이 빈번한 상황에서 데이타의 갱신을 효율적으로 처리하기 위한 색인 기법을 제안한다. 제안하는 색인구조는 대표적인 공간 분할 색인 기법 중 하나인 KDB-트리를 기반으로 하고 있으며, 캐시의 활용도를 높이기 위한 데이타 압축 기법과 포인터 제거 기법을 제안한다. 제안하는 기법의 우수성을 보이기 위해서 기존의 대표적인 캐시를 고려한 색인 구조중 하나인 CR-트리와 실험을 통해 성능을 비교하였으며, 성능평가 결과, 제안하는 색인 구조는 삽입 성능과 갱신 성능, 캐시 활용도 면에서 기존 색인 기법에 비해 각각 85%, 97%, 86% 의 성능이 향상되었다.

Keywords

References

  1. Phil Bernstain. et al., 'The Asilomar report on database research,' Sigmod Record, 27(4), 1998
  2. Peter A. Boncz, et al., 'Database architecture optimized for the new bottleneck: Memory access,' Proceedings of the 25th VLDB Conference, pp.54-65, 1999
  3. Anastassia Ailamaki, David J. DeWitt, Mark D. Hill and David A. Wood, 'DBMSs On A Modern Processor: Where Does Time Go?,' Proceedings of the 25th VLDB Conference, pp.266-277, 1999
  4. Jun Rao and Kenneth A. Ross, 'Cache Conscious Indexing for Decision-Support in Main Memory,' Proceedings of the 25th VLDB Conference, pp.78-89, 1999
  5. Jun Rao and Kenneth A. Ross, 'Making B+-Trees Cache Conscious in Main Memory,' Proceedings of the ACM SIGMOD Conference, pp.475-486, 2000 https://doi.org/10.1145/335191.335449
  6. Kihong Kim, Sang K. Cha and Keunjoo Kwon, 'Optimizing Multidimensional Index Trees for Main Memory Access,' Proceedings of the ACM SIGMOD Conference, pp.139-150, 2001 https://doi.org/10.1145/376284.375679
  7. Kaushik Chakrabarti and Sharad Mehrotra, 'The Hybrid Tree: An Index Structure for High Dimensional Feature Spaces,' Proceedings of the International Conference on Data Engineering, pp. 440-447, 1999
  8. John T. Robinson, 'The K-D-B-Tree: A Search Structure for Large Multidimensional Dynamic Indexes,' Proceedings of the ACM SIGMOD Conference, pp.10-18, 1981
  9. R. Orlandic and B. Yu, 'Estimating the Probability of Overlap between Multi-imensional Rectangles in the Analysis of Spatial Structures,' Information Sciences, 2001
  10. A. Guttman, 'R-trees: A Dynamic Index Structure for Spatial Searching,' Proceedings of ACM SIGMOD Conference, pp.47-57, 1984
  11. Ratko Orlandic, Byunggu Yu, 'Implementing KDBTrees to Support High-Dimensional Data,' Proceedings of the International Database Engineering & Applications Symposium, IEEE, pp.58-67, 2001
  12. Byunggu Yu, Tomas Bailey, Ratko Orlandic, Jothi Somavaram, 'KDBKD-Tree: A Compact KDB-Tree Structure for Indexing Multidimensional Data,' Proceedings of the International Conference on Information Technology: Coding and Computing[Computers and Communications], IEEE, pp.676-680, 2003
  13. S.T. Leutenegger, 'Multi Dimensional Data Sets,' http://www.cs.du.edu/~leut/MultiDimData.html, 1996