Journal of Information Technology Services (한국IT서비스학회지)

Korea Society of IT Services (한국IT서비스학회)
권호 표지

Improving Performance of HPC Clusters by Including Non-Dedicated Nodes on a LAN

LAN상의 비전용 노드를 포함한 HPC 클러스터의 확장에 의한 성능 향상

  • Published : 2008.12.31
Download PDF
⟨ Previous Next ⟩

Abstract

Recently the number of Internet firms providing useful information like weather forecast data is growing. However most of such information is not prepared in accordance with customers' demand, resulting in relatively low customer satisfaction. To upgrade the service quality, it is recommended to devise a system for customers to get involved in the process of service production, which normally requires a huge investment on supporting computer systems like clusters. In this paper, as a way to cut down the budget for computer systems but to improve the performance, we extend the HPC cluster system to include other Internet servers working independently on the same LAN, to make use of their idle times. We also deal with some issues resulting from the extension, like the security problem and a possible deadlock caused by overload on some non-dedicated nodes. At the end, we apply the technique in the solution of some 2D grid problem.

Keywords

References

  1. 김영균, 오길호, "LAN 환경에서 유휴시간 예약에 기반한 PC Cluster 설계", 한국정보과학회 2003년도 가을 학술발표논문집, 제30권, 제2호(III), 2003
  2. 김진미 외, "차세대 컴퓨팅을 위한 가상화 기술", ETRI 전자통신동향분석, 제23권, 제4호(2008)
  3. 남기찬, 김용진, "서비스사이언스 관점에서 본IT 서비스산업의 발전과제", 소프트웨어진흥원 SW Insight 정책리포트, 2008
  4. 이규웅, "SAN 기반 클러스터 파일 시스템 $SANique^{TM}$의 성능평가 및 분석", 한국IT서비스학회지, 제7권, 제1호(2008), pp.195-203
  5. Bittman, T., "The Future of Server Virtualization", Gartner Research Note T-20-4339, 2003
  6. Fagg, G. E., E. Gabriel, Z. Chen, T. Angskun, G. Bosilca, A. Bukovsky, and J. J. Dongarra, "Fault tolerant communication library and applications for high performance computing", Proceedings of the Los Alamos Computer Science Institute Symposium 2003, Santa Fe, NM
  7. Gartner, Inc., "Gartner Identifies the Top 10 Strategic Technologies for 2009", http://www.gartner.com/it/page.jsp?id = 777212
  8. Kaufman, L., "Matrix methods for queuing problems", SIAM J. Sci. Stat. Comput., Vol. 4(1983), pp.525-552 https://doi.org/10.1137/0904037
  9. Sankaran, S., J. M. Squyres, B. Barrett, and A. Lumsdaine, "The LAM/MPI Checkpoint /Restart Framework：System-Initiated Checkpointing", International J. of High Perfornamce Computing Applications, Vol.19 (2005), pp.479-493 https://doi.org/10.1177/1094342005056139
  10. Sloan, J., High Performance Linux Clusters with OSCAR, Rocks, OpenMosix, and MPI, O'Reilly Media Inc. 2005
  11. Subramaniyan, R., V. Aggarwal, A. Jacobs, and A. George, "FEMPI：A Lightweight Fault-tolerant MPI for Embedded Cluster Systems", Proc. of International Conference on Embedded Systems and Applications (ESA), Las Vegas, NV, June 26-29, 2006
  12. Zaki, M., W. Li, and S. Parthasarathy, "Customized Dynamic Load Balancing in a Heterogeneous Network of Workstations", In 5th IEEE Int. Symposium on High Perfornamce Distributed Computing, 1996
  13. Berkeley Lab Checkpoint/Restart(BLCR), http://ftg.lbl.gov/CheckpointRestart/CheckpointRestart.shtml
  14. Berkeley NOW Project, http://now.cs.berkeley.edu/
  15. Open MPI, http://www.open-mpi.org
  16. PCs by day, supercomputer by night, http://www.theglobeandmail.com/servlet/story/RTGAM.20060615.tqsuperjun15/BNStory/GlobeTQ
  17. http://www.linuxquestions.org/questions/linux-security-4/firewall-blocking-nfs-eventhough-ports-are-open-294069/
  18. http://www.vmware.com/kr/overview/
  19. http://kr.blog.yahoo.com/thisrule1/857793.html