Journal of KIISE:Computer Systems and Theory (한국정보과학회논문지:시스템및이론)

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

VIA-Based PC Cluster System for Efficient Information Retrieval

효율적인 정보 검색을 위한 VIA 기반 PC 클러스터 시스템

  • 강나영 (삼성전자 무선사업부) ;
  • 정상화 (부산대학교 컴퓨터공학과) ;
  • 장한국 (부산대학교 컴퓨터공학과)
  • Published : 2002.10.01
Download PDF
⟨ Previous Next ⟩

Abstract

PC cluster-based Information Retrieval (IR) systems improve their performances by parallel processing of query terms using cluster nodes. However TCP/IP based communication used to exchange data between cluster nodes prevents the performance from being improved further. The user-level communication mechanisms solve the problem by eliminating the time-consuming kernel access in exchanging data between cluster nodes. The Virtual Interface Architecture (VIA) is one of the representative user-level communication mechanisms which provide low latency and high bandwidth. In this paper, we propose a VIA-based parallel IR system on a PC cluster. The IR system is implemented using the following three communication methods: Sealable Coherent Interface (SCI) based VIA, MPI on SCI based VIA, MPI on Fast Ethernet based VIA. Through experiments, the performances of the three methods are analyzed in various aspects.

PC클러스터 기반 정보 검색 시스템은 질의를 클러스터 상의 노드에 분산시켜 병렬로 처리함으로써 전체 시스템의 성능을 향상시킬 수 있다. 그러나, 노드 사이의 데이터 교환을 위하여 TCP/IP 기반 통신을 사용하는 것은 전체 시스템 성능 저하의 원인이 된다. 이를 해결하기 위해 개발된 것이 사용자 수준 통신(user-level communication)이다. 이것은 성능에 치명적인 영향을 미치는 커널 접근을 통신 단계에서 제거함으로써 적은 지연시간과 높은 대역폭을 제공한다. 본 논문에서는 사용자 수준 통신 방법의 업계 표준인 VIA(Virtual Interface Architecture)를 기반으로 한 효율적인 병렬 정보 검색 시스템을 제안한다. 본 논문의 정보 검색 시스템은 SCI(Scalable Coherent Interface) 기반의 VIA 방식, SCI 기반의 VIA/MPI 방식 그리고 Fast Ethernet 기반의 VIA/MPI 방식으로 구현되었으며 실험을 통하여 세 방식의 성능을 비교 분석하였다.

Keywords

References

  1. Microprocessor and Microcomputer Standards Subcommittee, 'IEEE Standard for Scalable Coherent Interface,' IEEE Std 1596-1992, IEEE Computer Society, August 1993
  2. Basu, A., Buch, V., Vogels, W. and von Eiken, T., 'UrNet: A User-Level Network Interface for Parallel and Distributed Computing,' Proceeding of the 15th ACM Symposium on Operating Systems Principles, pp. 40-53, Copper Mountain, Colorado, United States, December 1995 https://doi.org/10.1145/224056.224061
  3. Blumrich, M., Dubnichi, C., Felten, E. W. and Li, K., 'Virtual Memory-Mapped Network Interfaces,' IEEE Micro, pp. 21-28, February 1995 https://doi.org/10.1109/40.342014
  4. Mainwaring, A. and Culler, C., 'Active Message Applications Programming Interface and Communication Subsystem Organization,' Technical Document, 1995
  5. Pakin, S., Kararncheti, V. and Chien, A. A., 'Fast Messages (FM): Efficient, Portable Communication for Workstation Clusters and Massively-Parallel Processors,' IEEE Concurrency, Vol. 5, No.2, pp. 60-73, 1997 https://doi.org/10.1109/4434.588295
  6. http://www.nersc.gov/research/FTG/via
  7. Sharma, R., 'A Generic Machine for Parallel Information Retrieval,' Information Processing and Management, Vol. 25, No.3, pp. 223-235, 1989 https://doi.org/10.1016/0306-4573(89)90041-1
  8. Cringean, J. K., England, R., Manson, G. A. and Willett, P., 'Network Design for the Implementation of Text Searching Using a Multicomputer,' Information Processing & Management, Vol. 27, No.4, pp. 265-283, 1991 https://doi.org/10.1016/0306-4573(91)90084-Y
  9. Stanfill, C and Thau, R., 'Information Retrieval on the Connection Machine: 1 to 8192 Gigabytes,' Information processing & Management, Vol. 27, No.4, pp. 285-310, 1991 https://doi.org/10.1016/0306-4573(91)90085-Z
  10. Chung, S., Kwon, B., Ryu, K., Jang, B., Kim, J, and Choi, C, 'Parallel Information Retrieval on an SCI-Based PC-NOW,' Lecture Notes in Computer Science, 1800, pp. 81-90, May 2000
  11. Chung, S., Kwon, H., Ryu, K., Chung, Y., jang, B. and Choi, C, 'Information Retrieval on an SCI-Based PC Cluster,' Journal of Supercomputing, Vol. 19, Issue 3, pp. 251-265, July 2001 https://doi.org/10.1023/A:1011178530932
  12. http://www.emulex.com/products/vi/clan1000.html
  13. http://www.qlogic.com/products/sanblade/sanblade_ 2300.asp
  14. Trams, M., Schlosser, R. and Rehm, W., 'Design Choices and First Results of Our VIA-Capable PCI-SCI Bridge,' Proceedings of CLUSTER 2000, pp. 349-350, Chemnitz, Germany, November 2000
  15. Buonadonna, P., Geweke, A. and Culler, A., 'An Implementation and Analysis of the Virtual Interface Architecture,' Proceedings of SC98, Orlando, Florida, United States, November 1998 https://doi.org/10.1109/SC.1998.10052
  16. Giacomini, F., Amundsen, T., Bogaerts, A., Hauser, R., Johnsen, B. D., Kohmann, B., Nordstrom, R. and Werner, P., 'Esprit Project 23174 - Software Infrastructure for SCI (SISC/), Version 2.1.1,' White Paper, Dolphin Interconnect Solutions, 1999
  17. http://www.nersc.gov/research/FTG/mvich/index.html
  18. Shin, J., Chung, S. and Hahn, W., 'An SCI-based Software VIA System for PC Clustering,' 2001 IEEE International Conference on Cluster Computing, Newport Beach, United States, October 2001 https://doi.org/10.1109/CLUSTR.2001.959982