This paper applies the traditional read-only optimization in the one and two phase protocols of commit processing in the context of pure main memory databasesystems. Thereby there is no need for logging log records t...
详细信息
ISBN:
(纸本)9781467356343
This paper applies the traditional read-only optimization in the one and two phase protocols of commit processing in the context of pure main memory databasesystems. Thereby there is no need for logging log records to backup node in read-only transactions. Through this optimization, the response time will be decreased in these transactions. Furthermore, the message complexity will be decreased which leads to the decrease of communicative traffic overhead during commit processing in distributedsystems. This decrease in multi-level distributed transactions which are the most common types of transactions is more observable. Also, a new read-only optimization is proposed about two phase protocols that remove both the commit processing phases for read-only transactions and also removes all the logging activities to backup node for partially read-only transactions. As a result, response times, the message and log complexity, and transactions concurrency level in comparison to traditional read-only optimization will be improved more.
Performance of OLAP queries becomes a critical issue as the amount of data in the data warehouses increases rapidly. To solve this performance issue, we proposed a high performance database cluster system called Hyper...
详细信息
ISBN:
(纸本)9783642241055;9783642241062
Performance of OLAP queries becomes a critical issue as the amount of data in the data warehouses increases rapidly. To solve this performance issue, we proposed a high performance database cluster system called HyperDB in which many PCs can be mobilized for excellent performance. In HyperDB, an OLAP query can be decomposed into sub-queries, and each of the sub-queries can be processed independently on a PC in a short time. But if an OLAP query has nested form (i.e., nested SQL), it could not be decomposed into sub-queries. In this paper, we propose a parallel distributed query processing algorithm for nested queries in HyperDB system. Traditionally, parallel distributed processing of nested queries is known as a difficult problem in the database area.
Secret sharing is a method for distributing a secret among a party of participants. Each of them is allocated a share of the secret, and the secret can only be reconstructed when the shares are combined together. We h...
详细信息
Secret sharing is a method for distributing a secret among a party of participants. Each of them is allocated a share of the secret, and the secret can only be reconstructed when the shares are combined together. We have been proposing a secret sharing distributed database system (SSDDB) that uses a secret sharing scheme to improve confidentiality and robustness of distributed database systems. This paper proposes a vertical partitioning algorithm for the SSDDB, and evaluates the algorithm by computational experiments.
There are the trends that the users hope to retrieve information among lots of individual databasesystems distributed in lots of virtual organizations. To connect all these individual databasesystems into a virtual ...
详细信息
There are the trends that the users hope to retrieve information among lots of individual databasesystems distributed in lots of virtual organizations. To connect all these individual databasesystems into a virtual databasesystem is a significant issue. This paper presents VIRGO_DDBMS, a framework of distributed database system based on virtual hierarchical tree Grid organizationsfVIRGO) P2P network. The table space are classified as hierarchical domains like DNS. The servers hosting databasesystems joins several groups of VIRGO network according to the owned tables' domains. VIRGO_DDBMS is effective for data update and retrieve by using distributed SQL statements. We use the policies which cache the host machines' addresses of databasesystems to avoid the overload of traffic of root nodes in tree structure. The proposal presented here can construct a virtual distributed database system which can manages huge volumes of data by connecting all databasesystems together which are owned by the individual virtual organizations.
The important problem of distributed database systems (DDBs) is "data allocation". There are many methods for this problem and there are two measures for compare these models: Minimal cost and Performance. I...
详细信息
ISBN:
(纸本)9780769535579
The important problem of distributed database systems (DDBs) is "data allocation". There are many methods for this problem and there are two measures for compare these models: Minimal cost and Performance. In this paper we use a new method for using genetic algorithm. At first we generate clusters based on the communication cost between the sites, then perform genetic algorithm on these clusters to rind which one is the best situation for allocate data, at last allocate data to their sites in the same way. We improve performance of DDB with reduce number of communication cost, data redundancy, and minimize total data transfer cost by using genetic algorithm and grouping sites. We compare our results with another model;our proposal model has a higher performance.
The Performance and the efficiency of a distributed database system depend highly on the way data are allocated to the sites. The NP-completeness of the data allocation problem and the large size of its real occurrenc...
详细信息
The Performance and the efficiency of a distributed database system depend highly on the way data are allocated to the sites. The NP-completeness of the data allocation problem and the large size of its real occurrence, call for employing a fast and scalable heuristic algorithm. In this paper, we address the data allocation problem in terms of minimizing two different types of data transmission across the network, i.e., data transmissions due to site-fragment dependencies and those caused by inter-fragment dependencies. We propose a new heuristic algorithm which is based on the ant colony optimization meta-heuristic, with regards to the applied strategies for query optimization and integrity enforcement. The goal is to design an efficient data allocation scheme to minimize the total transaction response time under memory capacity constraints of the sites. Experimental tests indicate that our algorithm is capable of producing near- optimal solutions within a reasonable time. The results also reveal the flexibility and scalability of the proposed algorithm.
Motivated by a problem faced by a multimedia entertainment retailer, we explore the problem of planning the design of a distributed database system. The problem consists of planning the design/expansion of the distrib...
详细信息
Motivated by a problem faced by a multimedia entertainment retailer, we explore the problem of planning the design of a distributed database system. The problem consists of planning the design/expansion of the distributed database system by introducing new database servers and retiring possibly some existing ones in order to reduce telecommunication costs for processing user queries and server acquisition, operations and maintenance cost in a multiperiod environment where user processing demand varies over time. We develop a mathematical programming model and an effective solution approach to determine the best decisions regarding acquisition and retirement of database servers and assignment of user processing demand to the servers over time. Through a computational study, we investigate the impact of important parameters such as length of the planning horizon and demand growth on the solution quality and utilization of server capacity and examine the effectiveness of the solution approach in comparison with the commercial package LINDO. We also discuss some extensions to the problem as directions for future research. (C) 2003 Elsevier Inc. All rights reserved.
The administration of distributed data is a need that nowadays many enterprises strive to fulfill. However, it is often difficult for these enterprises to find a software product capable of providing the solution they...
详细信息
The administration of distributed data is a need that nowadays many enterprises strive to fulfill. However, it is often difficult for these enterprises to find a software product capable of providing the solution they require. This is due to the inherent complexity to distributed data access, which has caused the majority of database applications to provide distributed access to a centralized database as a simpler alternative to distributed database systems. This article demonstrates that it is possible to design a component-based reference architecture of a distributed database systems (D-DBS) with a 3-tiered client-server structure from the solution given to the data distribution problem through the implementation of a global conceptual schema. Therefore, this proposal reduces the design complexity of a D-DBS.
Majority of the research in multilevel secure database management systems (MLS/DBMS) focuses primarily on centralized databasesystems. However, with the demand for higher performance and higher availability, database...
详细信息
ISBN:
(纸本)9783540770855
Majority of the research in multilevel secure database management systems (MLS/DBMS) focuses primarily on centralized databasesystems. However, with the demand for higher performance and higher availability, databasesystems have moved from centralized to distributed architectures, and the research in multilevel secure distributeddatabase management systems (MLS/DDBMS) is gaining more and more prominence. Traditional transaction management protocols (i.e., concurrency control and commit protocols) are important components of databasesystems. The most important issues for these protocols in MLS databasesystem are the covert channel problem [2] and starvation of high security level transactions [10]. To address these problems, first we propose new correctness criteria for multilevel secure multiversion concurrency control protocol, called read-down conflict serializability. It is the extended definition of one-copy serial (or1-serial) that allows a transaction to read older versions, if necessary. If a concurrency control protocol allows transaction to read older versions, we can obtain better throughput and response time than the traditional multiversion concurrency control protocols. We show that multiversion schedule based upon proposed criteria is also one-copy serializable. Secondly, this paper proposes a secure multiversion concurrency control protocol for MLS/DDBMSs that is only free from covert channels but also do so without starving high security level transactions, in addition to ensure the proposed serializability. Further, in distributed database systems, an atomic commitment protocol is needed to terminate distributed transactions consistently. To meet MLS requirements and to avoid database inconsistencies 2PC commit protocol is also modified.
By analyzing the characteristics and situation of highway, a highway synthetical information management system and its designing project is proposed. The mixed mechanism of B/S architecture and three tier C/S architec...
详细信息
By analyzing the characteristics and situation of highway, a highway synthetical information management system and its designing project is proposed. The mixed mechanism of B/S architecture and three tier C/S architecture are adopted. Combining with the geography information system (GIS), the system network platform is build up by using ASP. NET. Then the distributed database system based on the mixed mechanism of B/S architecture and three-tier C/S architecture is build up with Delphi7. The application of the synthetical information management system is realized in highway of Chongqing.
暂无评论