• Journal of KIISE:Databases
• /
• v.29 no.6
• /
• pp.502-512
• /
• 2002

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.

• Journal of KIISE:Databases
• /
• v.29 no.6
• /
• pp.500-500
• /
• 2002

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.

• Journal of Internet Computing and Services
• /
• v.5 no.1
• /
• pp.51-57
• /
• 2004

Mobile Internet allows users to request critical information and receive swift responses at any places, but mobile users could suffer from unreliable and ill-timed services due to the characteristics of wireless media, One way that reduces possibility of the unsatisfactory services is data replication. Data Replica1ion, however, inevitably induces the overhead of maintaining replica consistency which requires more expensive synchronization mechanism. We propose a new replicated data management scheme in distributed mobile environment, In order to alleviate negative impact of synchronization message overhead in fault-prone mobile Internet environment, we devise a new replication control scheme called proxy quorum consensus (PQC), PQC minimizes the message overhead by coordinating quorum access activities by means of proxy mediated voting (PMV) which exploits reliable proxy hosts instead of unreliable mobile hosts in voting process, We also propose a simulation model to show the performance of PQC. Based on the results of the performance evaluation, we conclude that PQC scheme outperforms the traditional schemes.

• Proceedings of the Korean Information Science Society Conference
• /
• 2004.10c
• /
• pp.97-99
• /
• 2004

ONS나 E-mail 혹은 Instant Messaging등 메시지 애플리케이션에서 서버-클라이언트 모델은 scalability에 의한 한계가 있다. 이러한 문제를 해결할 수 있는 대안은 Peer-to-Peer 기반 정근방식이다. 하지만 이러한 환경에서는 기존의 P2P 시스템에 높은 데이터의 지속성과 Bounded Look-up Latency를 효과적으로 지원하는 방안이 필수적이다. 본 논문에서는 Replication을 통하여 Look-up Latency를 줄이기 위한 하나의 방법으로 Distributed k-ary Replication(DkR)을 제안한다. DkR은 source node와 root node 사이에 multi-path를 유지하고 Look-up Locality 특성을 이용하여 Replica를 배치함으로써 효율적으로 Look-up latency를 줄이게 된다. 마지막으로 DkR의 Replication 효과를 시뮬레이션을 통해서 확인을 해 보았다.

• ETRI Journal
• /
• v.42 no.3
• /
• pp.388-398
• /
• 2020

We herein propose a heuristic redundancy selection algorithm that combines resubmission, replication, and checkpointing redundancies to reduce the resiliency overhead in fault-tolerant workflow scheduling. The appropriate combination of these redundancies for workflow tasks is obtained in two consecutive phases. First, to compute the replication vector (number of task replicas), we apportion the set of provisioned resources among concurrently executing tasks according to their needs. Subsequently, we obtain the optimal checkpointing interval for each task as a function of the number of replicas and characteristics of tasks and computational environment. We formulate the problem of obtaining the optimal checkpointing interval for replicated tasks in situations where checkpoint files can be exchanged among computational resources. The results of our simulation experiments, on both randomly generated workflow graphs and real-world applications, demonstrated that both the proposed replication vector computation algorithm and the proposed checkpointing scheme reduced the resiliency overhead.

• The Journal of Korean Association of Computer Education
• /
• v.7 no.3
• /
• pp.35-43
• /
• 2004

Internet-based P2P computing with independent machines suffers from frequent disconnections and security threats caused by leaving, failure, network diversity, or anonymity of participated machines. Replication schemes of shared resources are used for solving these issues in many studies and implementations. We propose an operational replication scheme in P2P computing to share computing resources, and the scheme verifies the correctness of operation against faults and security threats. This verifications are carried out periodically on replicated and dependent working units without global message exchanges over the whole system. The verified working units are treated as checkpoints, and thus they could be put to practical use for fault-tolerance with rollback recovery.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2008.10a
• /
• pp.53-56
• /
• 2008

Recently, the distributed system is being evolved into a new paradigm, named cloud computing, which provides users with efficient computing resources and services from data centers. Cloud computing would reduce the potential danger of Grid computing which utilizes resource sharing by constructing centralized data center. In this paper, a new data replication scheme is proposed for Hadoop distributed file system by changing 1:1 data transmission to 1:N. The proposed scheme considerably reduced the data transmission delay comparing to the current mechanism.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2010.07a
• /
• pp.15-16
• /
• 2010

최근 이슈가 되고 있는 클라우드 컴퓨팅은 대용량의 데이터를 분산 저장하고 제공할 수 있는 클러스터 파일 시스템을 필요로 한다. 이러한 클러스터 파일 시스템은 높은 신뢰성과 고가용성을 보장하기 위해서 파일 복제 기법을 사용하고 있다. 가장 많이 쓰이고 있는 복제 기법은 전체-파일 복제 기법으로 높은 파일 가용성을 제공하지만 그만큼 스토리지 오버헤드가 크다는 단점이 있다. 또 다른 복제 기법으로는 LDPC 코드를 이용한 것으로 비교적 적은 스토리지 오버헤드를 가지면서 동시에 비슷한 수준의 파일 가용성을 제공한다. 따라서 본 논문에서는 클러스터 파일 시스템을 위한 LDPC 코드 복제 기법의 최적화 방법을 제안한다.

• Journal of Internet Technology
• /
• v.19 no.6
• /
• pp.1801-1810
• /
• 2018

In this study, an energy-aware data-replication is proposed to effectively support a mobile sink in a solar-powered wireless sensor network (WSN). By utilizing the redundant energy efficiently, the proposed scheme shares the gathered data among the cluster heads using a backbone network, in order to increase data-reliability. It also maintains a backup cluster head in each cluster to enhance topological resilience. The simulation result showed that, compared to conventional clustering techniques, the proposed scheme decreases the total amount of data loss from the mobile sink as well as saving its energy (by reducing its moving distance), while minimizing the unexpected blackout time of the sensor node.

• Proceedings of the Korean Operations and Management Science Society Conference
• /
• 1995.09a
• /
• pp.84-84
• /
• 1995