The KIPS Transactions:PartA (정보처리학회논문지A)

Korea Information Processing Society (한국정보처리학회)
권호 표지
DOI QR코드

DOI QR Code

MPICH-GP : An MPI Extension to Supporting Private IP Clusters in Grid Environments

MPICH-GP : 그리드 상에서 사설 IP 클러스터 지원을 위한 MPI 확장

  • 박금례 (삼성전자 정보통신총괄 무선사업부 WiBro S/W Lab.) ;
  • 윤현준 (서강대학교 컴퓨터학과) ;
  • 박성용 (서강대학교 컴퓨터학과) ;
  • 권오영 (한국기술교육대학교) ;
  • 권오경 (KISTI 슈퍼컴퓨팅센터)
  • Published : 2007.02.28
Download PDF
⟨ Previous Next ⟩

Abstract

MPICH-G2 is an MPI implementation to solve complex computational problems by utilizing geographically dispersed computing resources in grid environments. However, the computation nodes in MPICH-G2 are exposed to the external network due to the lack of supporting the private IP clusters, which raises the possibility of malicious security attacks. In order to address this problem, we propose MPICH-GP with a new relay scheme combining NAT(Network Address Translation) service and an user-level proxy. The proxy running on the front-end system of private IP clusters forwards the incoming connection requests to the systems inside the clusters. The outgoing connection requests out of the cluster are forwarded through the NAT service on the front-end system. Through the connection path between the pair of processes, the requested MPI jobs can be successfully executed in grid environments with various clusters including private IP clusters. By simulations, we show that the performance of MPICH-GP reaches over 80% of the performance of MPICH-G2, and over 95% in ease of using RANK management method.

그리드 네트워크에서 MPI를 사용하여 지리적으로 산재된 컴퓨팅 자원을 활용하여 복잡한 문제를 해결하기 위한 MPICH-G2는 사설 IP 클러스터를 지원하지 못한다는 단점을 가지고 있다. MPICH-G2의 경우에는 모든 계산노드가 외부에 노출되어 악의적인 보안 침해의 가능성이 높아지게 된다. MPICH-GP는 NAT와 프락시를 병용하여 사설 IP 클러스터의 문제를 해결한다. 사설 IP 클러스터의 프론트 노드에 프락시를 두고, 이를 통해 내부 계산 노드로의 통신 링크를 중계한다. 따라서 사설 IP 클러스터와 공인 IP 클러스터가 혼재된 네트워크에서도 적정한 경로를 설정하고 성공적으로 MPI 작업을 수행할 수 있다. 본 논문에서는 MPICH-GP의 성능을 기존의 MPICH-G2와 비교하였다. 그리드 환경에서 MPICH-GP는 MPICH-G2의 80% 이상의 성능을 보이며, RANK 관리기법을 적용한 경우는 95% 이상의 성능을 나타낸다.

Keywords

References

  1. I. Foster, C. Kesselman and S. Tuecke. 'The Anatomy of the grid : Enabling scalable virtual organizations'. International Journals of Supercomputing Applications, 15(3), 2001 https://doi.org/10.1177/109434200101500302
  2. I. Foster and C. Kesselman. 'The Grid: A Blueprint for a New Computing Infrastructure'. Morgan Kaufmann, 1998
  3. W. Gropp, E. Lusk, and A. Skjellum. 'Using MPI: Portable Parallel Programming with the Message Passing Interface'. MIT Press, 1995
  4. W. Gropp, E. Lusk, N. Doss, and A. Skjellum. 'A high-performance, protable implementation of the MPI message passing interface standard.' Parallel Computing, Volume 22, pp.789-828, 1996 https://doi.org/10.1016/0167-8191(96)00024-5
  5. I. Foster, J Geisler, W. Gropp, N. Karonis, E. Lusk, G. Thiruvathukal, and S. Tuecke. 'A wide-area implementation of the Message Passing Interface.' Parallel Computing, pp. 1735-1749, 1998 https://doi.org/10.1016/S0167-8191(98)00075-1
  6. N. Karonis, B. Toonen, I Foster, 'MPICH-G2: a Grid-enabled implementation of the Message Passing Interface', Journal of Parallel and Distributed Computing, Volume 63, pp.551-563, 1998 https://doi.org/10.1016/S0743-7315(03)00002-9
  7. P. Srisuresh and K. Egevang, 'Traditional IP Network Address 'Translatort(Traditional NAT)', RFC 3022. International Engineering Task Force, Jan, 2001
  8. M. Muller, M. Hess, E. Gabriel, 'Grid enabled MPI solutions for Clusters', In 3rd International Symposium on Cluster Computing and the Grid, pp.18-25, 2003 https://doi.org/10.1109/CCGRID.2003.1199348
  9. Globus I/O API, 'Globus-IO reference' http://www.globus.org/v1.1/io/ globus_io.html
  10. E. Gabriel, M. Resch, T. Beisel and R. Keller, 'Distributed computing in a heterogeneous computing environment', In Recent Advances in Parallel Virtual Machine and Message Passing Interface, Lecture Nodes in computer Science. Springer, 1497, pp,180-197, 1998 https://doi.org/10.1007/BFb0056574
  11. Y. Tanaka, M. Sato, M. Hirano, H. Nakada, and S. Sekiguchi. 'Performance evaluation of a firewall-complaint globus-based wide-area cluster system.' In Proceedings of the Ninth IEEE International Symposium on High Performance Distributed Computing, pp. 121-128. IEEE Computing Society, 2000
  12. I. foster and C. Kesselman, and S. Tuecke. 'The Nexus approach to intergrating multithreading and communication. Journal of Parallel and Distributed Computing', pp.70-82, 1996 https://doi.org/10.1006/jpdc.1996.0108
  13. Globus DUROC, http://www.globus.org/duroc/frames.html
  14. S. Choi, K. Park, S. Han, S. Park, 'An NAT-Based Communication Relay Scheme for Private- IP-enabled MPI over Grid Environments', International Conference on Computational Science 2004 (ICCS 2004), pp 499-502, 2004
  15. Pallas MPI Benchmarks, http://www.pallas.com/e/products/pmb/
  16. NAS Parallel Benchmarks, http://www.nas.nasa.gov/Software/NPB
  17. Ford Bryan, Srisuresh Pyda, Kegel Dan. 'Peer-to-Peer Communication Across Network Address Translators', USENlX Annual Technical Conference, 2005