A distributed heterogeneous database system is a collection of interconnected local databasesystems with diverse properties and local autonomy. In such a system, it is difficult to maintain consistency without compro...
详细信息
ISBN:
(纸本)0818627557
A distributed heterogeneous database system is a collection of interconnected local databasesystems with diverse properties and local autonomy. In such a system, it is difficult to maintain consistency without compromising local autonomy. In this paper, a global concurrency control protocol is proposed that maintains consistency without compromising local autonomy. The scheme is based on the observation that sufficient read operations attached to global transactions can prevent local schedules from diverging while keeping the database states intact. The protocol is simple and easy to implement. It does not require the knowledge about local concurrency control methods. Thus local concurrency control mechanisms can be changed without affecting the global concurrency control scheme.
The purpose of this article is to characterize the inter-process communication (IPC) overhead in the context of multidatabase operations. A quantitative analysis of the cost of distributed query is becoming increasing...
详细信息
ISBN:
(纸本)0818626607
The purpose of this article is to characterize the inter-process communication (IPC) overhead in the context of multidatabase operations. A quantitative analysis of the cost of distributed query is becoming increasingly important as operations across multiple databases are becoming more common. A global nested-loop (NL) join method is one of the least-complicated methods for joining data across data sources. One of the advantages of the algorithm is that it works even when an external operation is a method call whose input parameters, unlike a query, must be bound at invocation time.
databasesystems are increasing in importance as storage repositories for applications like CAD and AI. These applications demand efficient support for queries that are more complex than the common business queries th...
详细信息
ISBN:
(纸本)0818626607
databasesystems are increasing in importance as storage repositories for applications like CAD and AI. These applications demand efficient support for queries that are more complex than the common business queries that are supported well by most systems. More queries are also of an ad-hoc nature, limiting the utility of precompilation approaches. The need for fast response to make AI and CAD applications feasible also has further motivated research into techniques for exploiting resources in a distributed system, for improved performance. This paper explores query decomposition techniques for complex query processing in distributed database systems and exposes the deficiencies associated with current algorithms.
This paper attempts to document some of the shortcomings of the optimistic concurrency control (OCC) approach in supporting all the features expected in a full-function DBMS. Surprisingly, in spite of OCC having been ...
详细信息
ISBN:
(纸本)0818626607
This paper attempts to document some of the shortcomings of the optimistic concurrency control (OCC) approach in supporting all the features expected in a full-function DBMS. Surprisingly, in spite of OCC having been around for a long time and its performance having been studied in various contexts, no complete system design, let alone a full-blown implementation, exists, as far as the author knows. The problems with OCC relate to support for access paths (indexes, hash-based storage), partial rollbacks, nested transactions, fine-granularity (e.g., record-level) conflict checking, different isolation levels, distributed transactions, and so on. The global of this paper is to increase awareness of the implementation aspects of OCC amongst researchers and to initiate a debate about the practical utility of OCC.
While researchers strive to develop new systems to enhance the cooperative document editing process, many authors already collaborate, using existing text processing systems to produce papers and reports. Using these ...
详细信息
ISBN:
(纸本)0897915429
While researchers strive to develop new systems to enhance the cooperative document editing process, many authors already collaborate, using existing text processing systems to produce papers and reports. Using these tools, one of the most time-consuming and error-prone collaboration tasks is maintaining a consistent shared bibliography. We have designed and implemented the BIBDB system to simplify collaborative authoring by providing a shared, cooperatively maintained bibliographic database. BIBDB uses existing networking technology and merges seamlessly into the LATEX/BIBTEX text processing system. The contributions of BIBDB include a set of user interface policies and software implementation techniques that support cooperative database maintenance.
This paper describes query processing in the Paryrus integrated data server. Query processing is unique in Papyrus because the underlying language is very powerful (computationally complete), performance on uni-proces...
详细信息
ISBN:
(纸本)0818626607
This paper describes query processing in the Paryrus integrated data server. Query processing is unique in Papyrus because the underlying language is very powerful (computationally complete), performance on uni-processor systems is not sacrificed in order to support multiprocessor architectures, and queries are self-scheduling in order to support extensibility. A prototype demonstrating the feasibility of this approach is in operation.
Data consistency in distributeddatabase management systems is often implemented by transactions which separately preserve consistency. Concurrent transaction execution can be ensured by locks on data and the discipli...
详细信息
Data consistency in distributeddatabase management systems is often implemented by transactions which separately preserve consistency. Concurrent transaction execution can be ensured by locks on data and the discipline of two-phase locking. Transactions can be forced to wait for others to release locks. It is possible that a complete cycle of waiting transactions can form deadlock. In a distributed memory environment, some deadlock can be detected locally, i.e. without using the communications network. But the most difficult case to detect is when a number of transactions on more than one network node all wait for each other so that none can proceed. This is called distributed deadlock. We present two simple algorithms for the detection of distributed deadlock on a shared-nothing distributeddatabase system. They are designed to be scalable and minimize network message traffic. An object-oriented approach is used.
This paper proposes a permanent time-stamp ordering mechanism as a concurrency control mechanism in the distributeddatabase, where transactions are served in the order of their arrivals. In the proposed mechanism, th...
详细信息
This paper proposes a permanent time-stamp ordering mechanism as a concurrency control mechanism in the distributeddatabase, where transactions are served in the order of their arrivals. In the proposed mechanism, the read time stamp and the write time stamp are recorded on each data item;and when a conflict arises between transaction services, the transaction with a larger time stamp is aborted. This paper describes the algorithm for the permanent time-stamp ordering mechanism and confirms the validity of the algorithm.
Research in distributed artificial intelligence (DAI) has produced many architectures such as the blackboard model or more general decentralised system. Most of these models fall short in providing an open platform to...
详细信息
Research in distributed artificial intelligence (DAI) has produced many architectures such as the blackboard model or more general decentralised system. Most of these models fall short in providing an open platform to integrate various heterogeneous knowledge and data sources. Federated distributeddatabases are used to access multiple databases in a distributed system. The authors present a cooperative and intelligent architecture which has the advantage of providing an open platform to integrate various heterogeneous information and knowledge sources in a way that is flexible and able to adjust to changes. In the most general case it would integrate expert systems (ES), knowledge based systems (KBS), deductive and other databases, various heterogeneous data stores and 'the user'. The latter can be accessed by an intelligent human agent. The agents can be implemented in form of 'interfaces' to the distributedsystems. These interfaces are practically agents that know how to interact with humans.< >
暂无评论