A Maximum Mechanism of Data Transfer Rate using Parallel Transmission Technology on High Performance Network

고성능 네트워크에서 병렬 전송 기술을 이용한 전송률 극대화 메커니즘

  • 김영신 (경희대학교 전자정보학부) ;
  • 허의남 (경희대학교 전자정보학부)
  • Published : 2007.09.15

Abstract

Even though Internet backbone speeds have increased in the last few years due to projects like Internet 2 and NGI, many high performance distributed applications are able to achieve only a small fraction of the available bandwidth. The cause of such problem is due to a character of TCP/IP. The primary goal of this protocol is reliable data transmission. Therefore high speed data transmission didn't be considered when TCP/IP is designed. Hence several researchers have been studied in order to solve the problem of TCP/IP. One of these research results, parallel transfer technique, solves this problem to use parallel TCP connections on application level. Additionally, this technique is compatibility. Recently, these researchers have been studied a mechanism to decide the number of parallel TCP connections. However, some researchers reported the number of parallel TCP connection base on only empirical results. Although hardware performance of host affects transmission rate, the hardware performance didn't be considered in their works. Hence, we collect all data related to transmission rate, such as hardware state information (cpu utilization, interrupt, context switch). Then, we analyzed collected data. And, we suggest a new mechanism determining number of parallel TCP connections for maximization of performance based on our analysis.

NGI나 Internet2와 같은 프로젝트로 인해 인터넷 백본 속도가 상당히 높아졌음에도 불구하고, 분산된 응용 프로그램들은 고성능의 네트워크를 제대로 활용하지 못하고 있다. 이러한 현상이 발생하는 원인으로 표준 전송 프로토콜(TCP)을 들 수 있다. TCP는 안전성/신뢰성을 보장하기 위해 설계되어 있으나, 이로 인해 발생될 수 있는 성능 저하에 관한 문제는 고려되지 않았다. 이러한 문제를 해결하고자 여러 기술들이 연구되고 있으며, 그 중 병렬 전송 기술은 응용레벨에서 다중 스트림을 이용하여 데이타를 전송하는 기술로써, 호환성 문제까지 해결하고 있다. 최근 병렬 전송 기술을 연구하는 연구자들은 최적의 병렬연결 개수의 범위를 찾는데 연구의 초점을 맞추고 있다. 그러나 이러한 연구들에서는 최적의 병렬연결 개수를 실험을 통해 경험적으로 결정하고 있으며, 데이타를 전송하는 호스트의 성능이나 전송 거리는 고려하지 않고 있다. 따라서 본 논문에서는 호스트의 성능과 병렬 전송과의 관계, 전송 거리와 병렬 전송 관계를 분석하고, 그 결과를 토대로 효율적이면서 최대 전송 성능을 확보할 수 있는 최적의 병렬연결 개수 결정 메커니즘을 논의하고자 한다.

Keywords

References

  1. T. Dunigan, M. Mathis and B. Tierney, 'A TCP Tuning Daemon,' Proceeding of IEEE Supercomputing 2002 Conference, Baltimore, Maryland, LBNL-51022, Nov. 2002
  2. Yunhong Gu, Xinwei Hong, Marco Mazzucco, Robert Grossman, 'SABUL: A High Performance Data Transfer Protocol,' IEEE COMMUNICATIONS LETTERS, 2003
  3. Aaron falk, Ted Faber, Joseph Bannister, Andrew Chien, Robert Grossman, Jason Leigh, 'Transport Protocols for High Performance,' Communications of The ACM, Vol.46, No11, November 2003
  4. Eric Weigle, Wu-chun Feng, 'Dynamic Right- Sizing: A Simulation Study,' IEEE ICCCN, 2001
  5. J.Semke, J.Mahdavi, M.Mathis, Automatic TCP Buffer Tuning, ACM SIGCOMM 1998, Vol.28, No.4, 1998
  6. Gi-chul Yoo, Eun-sook Sim, Dongkyun Kim, Taeyoung Byun, Kookhan Kim, Okh- wan Byun, 'An Efficient TCP Buffer Tuning Technique Based on Packet Loss Ratio (TBT-PLR),' International Conference on Internet Computing, 2004
  7. White paper GridFTP Universal Data Transfer for the Grid, 2000
  8. Harimath Sivakumar, Stuart Bailey, Robert L.Grossman, 'PSockets: The Case for Applicationlevel Network Striping for Data Intensive Applications using High Speed Wide Area Networks,' SC2000: High-Performance Network and Computing, 2000
  9. Thomas J. Hacker, Brain D. Athey, 'The End-to-End Performance Effects of Parallel TCP Sockets on a Lossy Wide-Area Network,' Parallel and Distributed Processing Symposium (IPDPS), 2002
  10. Bill Allcock, Igor Mandrichenko, Timur Perelmutov, 'GridFTP v2 Protocol Description,' Germi National Accelerator Laboratory, 2004
  11. W. Allcock, J. Bester, J. Bresnahan, A. Chervenak, L. Liming, and S. Tuecke, 'GridFTP: Protocol Extensions to FTP for the Grid,' Argonne National Lab., April 2002
  12. Pawel Plaszczak, Joe Link, Rich Wellner, 'Grid- FTPexplained,' Paul Hubbard, http://www.icslab. agh.edu.pl/~kzajac/pa-wel/GridFTPexplained.doc
  13. Kun Myon Choi, Eui-Nam Huh, and Hyunseung Choo, 'Efficient Resource Management for TCP Parallel Streams with Buffer Tuning,' Springer- Verlag Lecture Notes in Computer Science, Vol. 3823, pp. 683-692, December 2005
  14. 'NISTnet A Linux based Network Emulation Tool,' http://snad.ncsl.nist.gov/nistnet/, retrieved on June 2003
  15. Young-Sin Kim, Eui-Nam Huh, 'Analysis of the Interference between Parallel Sockets Connections and Predictions of the Bandwidth,' International Journal of Computer Science and Network Security, 2005