한국정보과학회논문지:데이타베이스 (Journal of KIISE:Databases)

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

MOERS: 그룹 통신의 메시징 순서를 확장한 즉시 갱신 중복 기법

MOERS: An Eager Replication Scheme using Extended Messaging Order of Group Communication

  • 문애경 (한국전자통신연구원 이동분산처리팀) ;
  • 남궁한 (한국전자통신연구원) ;
  • 조행래 (영남대학교 전자정보공학부)
  • 발행 : 2002.12.01
PDF 다운로드
⟨ 이전 논문 다음 논문 ⟩

초록

기존에 제안된 대부분의 중복 기법은 원본 트랜잭션을 완료한 후 비동기적으로 중복사본에 대한 갱신 요구를 방송하는 지연 갱신 기법을 가정하였다. 지연 갱신 기법은 즉시 갱신 기법에서 발생하는 빈번한 교착상태 발생 문제를 해결할 수는 있지만 데이타 일관성 유지가 사용자 책임이라는 단점을 갖는다. 최근 그룹 통신의 메시징 순서를 적용한 즉시 갱신 중복 기법들이 제안되고 있다. 이들 기법은 메시징 순서를 이용하여 교착상태 발생률을 줄였지만, 송신 노드는 갱신 요구 메시지를 방송한 후 전역 직렬성을 검증하는 낙관적 기법을 채택하기 때문에 동시성이 증가할수록 철회 트랜잭션의 실행 오버헤드가 증가한다는 문제점을 갖는다. 본 논문에서는 철회 트랜잭션의 갱신 매시지 방송과 실행 오버헤드를 줄일 수 있는 즉시 갱신 중복 기법을 제안한다. 제안한 기법은 갱신 요구 매시지를 방송하기 전에 전역 직렬성 검사가 이루어지기 때문에 완료 트랜잭션은 한번의 메시지 방송으로 처리할 수 있다. 뿐만 아니라, 철회 트랜잭션은 다른 노드로 방송할 필요가 없기 때문에 메시지 전송 횟수를 줄일 수 있으며, 철회 트랜잭션의 실행으로 인한 디스크 액세스 수와 로크 대기 시간을 줄임으로써 성능을 향상시킨다.

Most of previous replication schemes assume lazy schemes, which asynchronously propagate replica updates to other nodes after the updating transaction commits. While lazy schemes are novel in the sense that they can resolve high deadlock frequencies of eager schemes, they suffer from the fact that users should be responsible for the data consistency. Recently replication schemes based on the messaging order of group communication have been studied. They can reduce the deadlock frequencies using the messaging order, but they have another problem of increasing execution overhead of aborted transactions at high degree of concurrency. This is because the sender node validates global serializability after multicasting replica update messages to other nodes. In this paper, we propose an efficient eager replication scheme for reducing the number of messaging and overhead of aborted transactions significantly. The proposed scheme validates any conflicts before multicasting and does not multicast when there is any conflict; hence, it can exploit performance improvement by alleviating the message traffic for replica update or by reducing the number of disk accesses and locking delay due to aborted transactions.

키워드

참고문헌

  1. Gray, P. Helland, P. O'Neil and D. Shasha, 'The Dangers of Replication and a Solution,' Proc. ACM SIGMOD, pp. 173-182, 1997 https://doi.org/10.1145/233269.233330
  2. T. Anderson, Y. Breitbart, H.F. Korth and A. Wool, 'Replication, Consistency, and Practicality: are These Mutually Exclusive?' Proc. ACM SIGMOD, pp. 484-495, 1998 https://doi.org/10.1145/276304.276347
  3. Y. Breitbart and H.F. Korth, 'Replication and Consistency: being Lazy Helps Sometimes,' Proc. 16th ACM Symposium on Principles of Database Syst., pp. 173-184, 1997 https://doi.org/10.1145/263661.263681
  4. P. Chundi, D.J. Rosenkrantz and S. Ravi, 'Deferred Updates and Data Placement in Distributed Databases,' Proc. Int'l Conf. on Data Eng., pp. 469-476, 1996 https://doi.org/10.1109/ICDE.1996.492196
  5. E. Pacitti, P. Minet and E. Simon, 'Fast Algorithms for Maintaining Replica Consistency in Lazy Master Replicated Databases,' Proc. Int'l Conf. on VLDB, pp. 126-137, 1999
  6. M. Butrico et al., 'Gold Rush: Mobile Transaction Middleware with Java-Object Replication,' Proc. 3rd USENIX Conf. on Object-Oriented Technologies and Syst., pp. 91-101, 1997
  7. J. Holliday, D. Agrawal and A. Abbadi, 'The Performance of Database Replication with Group Multicast,' Proc. IEEE 29th Int'l Symposium on Fault Tolerant Computing, pp. 158-165, 1999 https://doi.org/10.1109/FTCS.1999.781046
  8. J. Holliday, D. Agrawal and A. Abbadi, 'Using Multicast Communication to Reduce Deadlock in Replicated Databases,' Proc. 19th IEEE Symposium on Reliable Distributed Syst., pp, 196-205, 2000 https://doi.org/10.1109/RELDI.2000.885407
  9. B. Kemme and G. Alonso, 'A Suite of Database Replication Protocols based on Group Communication Primitives Distributed Computing Systems,' Proc. 18th Int'l Con! on Distributed Computing Syst., pp. 156-163, 1998 https://doi.org/10.1109/ICDCS.1998.679498
  10. B. Kemme and G. Alonso, 'Don't be Lazy, be Consistent: Postgres-R, A New Way to Implement Database Replication,' Proc. Int'l Conf. on VLDB. pp. 134-143, 2000
  11. B. Kemme and G. Alonso, 'A New Approach to Developing and Implementing Eager Database Replication Protocols,' A CM Trans. Database Syst., Vol. 25, No.3, pp. 333-379, 2000 https://doi.org/10.1145/363951.363955
  12. B. Kemme, F. Pedone, G. Alonso and A. Schiper, 'Processing Transactions over Optimistic Atomic Broadcast Protocols,' Proc. 19th IEEE Int'l Conf. on Distributed Computing Syst., pp. 424-431, 1999 https://doi.org/10.1109/ICDCS.1999.776544
  13. M. Carey, M. Franklin and M. Zaharioudakis, 'Fine-Grained Sharing in Page Server DBMS,' Proc. ACM SIGMOD, pp. 359-370, 1994 https://doi.org/10.1145/191839.191911
  14. S. Mishra and L. Wu, 'An Evaluation of Flow Control in Group Communication,' IEEE/ACM Trans. Networking, Vol. 6, No. 5, pp. 571-587, 1998 https://doi.org/10.1109/90.731193
  15. M.F. Kaashoek and A. Tanenbaum, 'An Evaluation of the Amoeba Group Communication System,' Proc. 16th Int'l Conf. on Distributed Computing Syst., pp. 436-447, 1996 https://doi.org/10.1109/ICDCS.1996.507992
  16. K.P. Birman, A. Schiper, and P. Stephenson, 'Lightweight Casual and Atomic Group Multicast,' ACM Trans. Computer Syst., Vol. 9, No.3, pp. 272-314, 1991 https://doi.org/10.1145/128738.128742
  17. L. Moser et al., 'Totem: A Fault-tolerant Multicat Group Communication System,' Comm. ACM, Vol. 39, No. 4, pp. 54-63, 1996 https://doi.org/10.1145/227210.227226
  18. H. Schwtman, CSIM User Guide for use with CSIM Revision 16, MCC, 1992
  19. H. Cho, 'Performance of Cache Coherency Schemes in a Shared Disks Transaction Environment,' Proc. 6th IEEE Workshop on Future Trends of Distributed Computing Syst., pp. 154-161, 1997 https://doi.org/10.1109/FTDCS.1997.644718
  20. Y.C. Tay, N. Goodman and R. Suri, 'Locking Performance in Centralized Databases,' ACM Trans. Database Syst., Vol. 10, No. 4, pp. 415-462, 1085 https://doi.org/10.1145/4879.4880