http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
- 中文 을 입력하시려면 zhongwen을 입력하시고 space를누르시면됩니다.
- 北京 을 입력하시려면 beijing을 입력하시고 space를 누르시면 됩니다.
RISS 인기검색어
- 검색키워드
- 저자 : 박정량(Jeong Ryang Park) (검색결과 2 건)
- 무료
- 기관 내 무료
- 유료
-
-
데이타베이스 공유 환경에서 분산 동시성 제어를 위한 캐쉬 일관성 기법
김신희,류명춘,박정량 한국정보시스템학회 1998 情報시스템硏究 Vol.7 No.2
In Database sharing system, since multiple nodes may be simultaneously cached a page, cache coherency must be ensured so that every node can always access the latest version of pages. In this paper, we propose efficient cache coherency schemes in DBSS, where the database is logically partitioned using primary copy authority to reduce locking overhead. The proposed schemes can improve performance by reducing the disk I/O overhead and the message overhead due to maintaining cache coherency. Furthermore, they can show good performance when database workloads are varied dynamically. In this paper, we have proposed two cache coherency schemes such as DPCA_P and DPCA_U in the DBSS. The proposed schemes can reduce the disk I/O overhead and the communication overhead of conventional cache coherency scheme (SPCA) by reallocating PCA dynamically. Specifically, DPCA_P reallocates PCA when a page is replaced from the local buffer of PCA node. Hence, DPCA_P may alleviate the problem of excessive page faults of SPCA, since the replaced page need not be written to disk if another node caches the page. DPCA_U reallocates PCA when a node tries to update a page by requesting X lock to PCA node. This means that DPCA_U does not require the page transfer at transaction commit since an updating node becomes a new PCA node. We have explored the performance of DPCA_P and DPCA_U under a wide variety of database workloads and system configurations. The basic results obtained from the experiments can be summarized as follows. If the access pattern of transaction shows a low degree of locality of references, DPCA_P and DPCA_U outperform SPCA about 20% and 10%, respectively. However, the degree of performance improvements is reduced significantly when the locality of references is high. Note that providing good performance at low degree of locality of references is very promising, since in most. DBSS applications the degree of locality of references are about 30%. If the number of participating nodes are not large, DPCA_U performs better than SPCA even though all nodes have the same data access skew. However, as the number of nodes are increased, DPCA_U is outperformed by SPCA due to frequent PCA reallocations.
연관 검색어 추천
이 검색어로 많이 본 자료
활용도 높은 자료
