• Proceedings of the Korean Information Science Society Conference
• /
• 2002.04b
• /
• pp.211-213
• /
• 2002

데이터베이스의 활용도가 크게 증가되고 사용자가 증가되면서 데이터베이스의 가용성과 확장성이 중요시되고 있다. 이에 따라 확장 가능한 데이터베이스는 실시간 트랜잭션의 처리를 위해 온-라인 상태로 중단 없이 동작해야 한다. 사용자의 증가에 따른 질의의 집중 현상을 해결하기 위한 데이터베이스는 사용자의 질의를 처리하면서도 확장이 가능해야 한다. 또한, 온-라인 학장은 확장 가능한 고가용 데이터베이스에서 트랜잭션의 결과 응답 시간에 영향을 미치지 않고 트랜잭션의 처리량의 저하가 없어야 한다 본 논문에서는 질의 집중 현상을 해결하기 위해 기존의 데이터베이스에서 제안된 기법들을 살펴보고 온-라인 확장에 대한 기존 연구에서의 문제점을 보이며. 개선된 은-라인 확장 기법을 제안한다 제안되는 기법은 불필요한 확장 영역을 축소시키고 확장되는 노드에 대한 정책을 변형하여 내부 네트웍 사용을 줄임과 동시에, 데이터 복사의 병렬성을 향상시킨다. 본 연구를 통해 개선된 확장기법은 온-라인 확장 시 데이터베이스의 처리량과 트랜잭션 응답 속도를 향상시키고 확장성을 유지한다

• The Transactions of the Korea Information Processing Society
• /
• v.5 no.11
• /
• pp.2751-2762
• /
• 1998

In this paper, a new priority assignment ploicy and concurrency control for improvement of transaction predictability and performance are proposed. We present a new priority assignment algorithm called classified priority assignment(CPA), which solves the defects of Earliest Deadline First(EDF) by using class and bucket, and deals with real-time transaction and time-sharing transaction effectively. Also, we present a new concurrency control policy called conditional optimistic concurrency control using lock. It uses optimistic concurrency control for improvement of predictability, and solves transaction conflict by considering priority and execution time of transaction to waste less system resource.

• KIISE Transactions on Computing Practices
• /
• v.23 no.8
• /
• pp.466-472
• /
• 2017

Hardware Transactional Memory (HTM) has greatly changed the parallel programming paradigm for transaction processing. Since Intel has recently proposed Transactional Synchronization Extension (TSX), a number of studies based on HTM have been conducted. However, the existing studies support conflict prediction for a single cause of the transaction processing and provide a standardized TSX environment for all workloads. To solve the problems, we propose an efficient hardware transactional memory scheme for processing transactions in multi-core in-memory environment. First, the proposed scheme determines whether to use Software Transactional Memory (STM) or the serial execution as a fallback path of HTM by using a prediction matrix to collect the information of previously executed transactions. Second, the proposed scheme performs efficient transaction processing according to the characteristic of a given workload by providing a retry policy based on machine learning algorithms. Finally, through the experimental performance evaluation using Stanford transactional applications for multi-processing (STAMP), the proposed scheme shows 10~20% better performance than the existing schemes.

• The KIPS Transactions:PartC
• /
• v.10C no.3
• /
• pp.325-334
• /
• 2003

In stateless server, if an asynchronous cache invalidation scheme attempts to support local processing of read-only transaction in mobile client/sever database systems, a critical problem may occur ; the asynchronous invalidation reports provide no guarantees of waiting time for mobile transactions requesting commit. To solve this problem, the server in our algorithm broadcasts two kind of messages, asynchronous invalidation report to reduce transaction latency and periodic guide message to avoid the uncertainty of waiting time for the next invalidation report. The asynchronous invalidation report has its own sequence number and the periodic guide message has the sequence number of the most recently broadcast asynchronous invalidation report. A mobile client checks its cache validity by using the sequence numbers of these messages.

• Proceedings of the Korea Database Society Conference
• /
• 1997.10a
• /
• pp.504-514
• /
• 1997

관계형 데이터베이스의 성능을 향상시키는데 중요한 요소 중의 하나는 트랜잭션을 처리하기 위해 데이터를 디스크에서 주기억장치로 옮기는데 필요한 디스크 액세스(access) 횟수이다. 본 연구는 관계형 데이터베이스에서 트랜잭션을 처리할 때, 릴레이션(relation)을 수직분할하여 디스크에 단편(fragment)으로 저장하므로써 필요한 단편만 액세스하여 액세스 횟수를 감소시키는데 목적이 있다. 단편에 속성을 중복할당하여 수직분할하므로써 트랜잭션을 만족시키는 단편의 수를 감소시켜 중복할당을 고려하지 않은 방법보다 디스크 액세스 횟수를 감소시킬 수 있다. 갱신트랜잭션의 경우 하나의 속성이 갱신되면 중복된 속성을 모두 갱신하여야 하므로 액세스 횟수가 증가하지만, 조회트랜잭션의 경우 각 단편에 속성을 중복할당하여 액세스 횟수를 감소시킬 수 있다. 본 연구에서는 속성의 중복을 허용하여 단편을 구성하는 경우에 중복을 고려하지 않은 경우를 포함하므로 효과적으로 디스크 액세스 횟수를 감소시킬 수 있다. 본 연구에서는 중복할당을 고려하여 디스크의 액세스 횟수를 최소화시킬 수 있는 수직분할문제의 0-1 정수계획모형을 개발하고, 모형에 대한 최적해법으로 분지한계법을 제안한다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2000.04a
• /
• pp.61-66
• /
• 2000

본 논문에서는 분산 실시간 데이터베이스 시스템을 위한 효율적인 동시성 제어 기법을 제안한다. 제안하는 기법은 분산 실시간 환경에서 완료준비 단계에 도달한 트랜잭션의 우선 순위를 상승시킴으로써 트랜잭션의 재시작에 의한 낭비를 줄이고, 트랜잭션의 완료를 최대한 보장하며, 잠금 지연 시간을 최소화하는 장점을 갖는다. 또한, 제안하는 기법은 우선 순위가 상승된 트랜잭션의 완료를 보장하며 데이터 차용(borrowing)을 통해 다른 트랜잭션의 지연시간을 줄여줌으로써 전체적인 시스템 성능을 향상시킨다.

• The KIPS Transactions:PartD
• /
• v.8D no.1
• /
• pp.10-23
• /
• 2001

Transactions in real-time active databases have the notion of activeness where transactions are generated by external effects and another transaction. In this paper, we propose the multi version concurrency control algorithm for real-time active transactions. A real-time active transaction has a timing constraint in the form of a deadline until which the user wants to complete the transaction, and is characterized by triggering relationships which mean that association between a transaction that triggers the execution of another transaction and the triggered transaction. The triggering relationship is an important factor to resolve data conflicts among real-time active transactions. The proposed concurrency control mechanism resolves data conflicts by considering triggering relationships between conflicting transactions as well as priorities and precedence relationships. The conflict resolution mechanism considers association types of the triggering relationship such as abort and commit dependency, and then resolves data conflicts in favor of higher priority transactions. We also present the experimental results of our algorithm comparing other real-time active concurrency control algorithms.

• Journal of KIISE:Databases
• /
• v.30 no.6
• /
• pp.629-640
• /
• 2003

A shared disk (SD) cluster couples multiple nodes for high performance transaction processing, and all the coupled nodes share a common database at the disk level. In the SD cluster, a transaction routing corresponds to select a node for an incoming transaction to be executed. An affinity-based routing can increase local buffer hit ratio of each node by clustering transactions referencing similar data to be executed on the same node. However, the affinity-based routing is very much non-adaptive to the changes in the system load, and thus a specific node will be overloaded if transactions in some class are congested. In this paper, we propose a dynamic transaction routing scheme that can achieve an optimal balance between affinity-based routing and dynamic load balancing of all the nodes in the SD cluster. The proposed scheme is novel in the sense that it can improve the system performance by increasing the local buffer hit ratio and reducing the buffer invalidation overhead.

• The KIPS Transactions:PartD
• /
• v.13D no.1 s.104
• /
• pp.139-146
• /
• 2006

This paper describes the design and implementation details of a distributed transactional workflow monitoring system. There have been prevalent research and development trends in the workflow literature - workflow systems tend to be completely distributed architectures to support very large-scale workflow applications on object-oriented and internet-based infrastructures. That is, the active (object), distributed (architecture), system-oriented (transaction), and large-scale (application) workflow systems are the key targets in terms of the research and development aspects. While the passive, centralized, human-oriented, and small/medium scale workflow systems are the typical instances of the traditional workflow systems. Unlike in the traditional (the client-server architecture) workflow systems, the workflow monitoring features should not be easily supported in the recent (the fully distributed architecture) workflow systems. At the same time, they need a set of additional monitoring features, such as gathering and displaying statistical (or overload status) information of the workflow architectural components dispersed on the internet. We, in this paper, introduce the additional workflow monitoring features that are necessarily required for the recent workflow systems, and show how to embed those features into a web-based distributed workflow system.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2008.11a
• /
• pp.336-339
• /
• 2008

우리가 사용하는 데이터베이스 내에는 많은 양의 데이터 들이 들어 있으며, 계속적으로 그 양은 늘어나고 있다. 이러한 데이터들로부터 질의를 통해 얻을 수 있는 기본적이고 단순한 정보들과 달리 고급 정보를 얻게 해주는 방법이 데이터 마이닝이다. 데이터 마이닝의 기법 중에서 본 논문에서는 k-means 알고리즘을 사용하여 트랜잭션을 클러스터링 함으로써 데이터베이스의 트랜잭션 수를 줄여 연관규칙의 대표적인 알고리즘인 Apriori 알고리즘의 단점인 트랜잭션 스캔으로 인한 성능 저하를 개선하고자 한다.