The transaction model provides a good solution to many programming problems, but has been recognized as too strict for many applications. Different parallel and distributed applications have different consistency requ...
详细信息
The transaction model provides a good solution to many programming problems, but has been recognized as too strict for many applications. Different parallel and distributed applications have different consistency requirements, so multiple concurrency control policies are needed. When data is shared among applications with different policies, then the policies must operate simultaneously and compatibly. The authors investigate the interoperability of transactions and another consistency model, atomic blocks, which provide exclusive access to an individual object. They explore the implications of 'uncontrolled' concurrency within the same COOPS. The work has been in the context of the MELD object-oriented programming language.< >
This conference proceedings contains 19 papers. The following topics are dealt with: recovery;replication;network architecture;reliable communication;performance analysis;evaluation and modeling;and simulation and tes...
详细信息
This conference proceedings contains 19 papers. The following topics are dealt with: recovery;replication;network architecture;reliable communication;performance analysis;evaluation and modeling;and simulation and testing.
A major obstacle in implementing a rollback recovery scheme for fault tolerance in a concurrent distributed system is the domino effect. A low overhead checkpointing scheme is proposed to prevent this effect. Each pro...
详细信息
A major obstacle in implementing a rollback recovery scheme for fault tolerance in a concurrent distributed system is the domino effect. A low overhead checkpointing scheme is proposed to prevent this effect. Each process saves its state periodically. The state-save synchronization among processes is implemented by bounding clock drifts. A communication protocol that assures that all saved states are consistent is developed.
Multicast communication in a distributed system connected by a local area network can increase parallelism, and it can also provide a greater functionality than one-to-one communication. In the authors' multicast ...
详细信息
Multicast communication in a distributed system connected by a local area network can increase parallelism, and it can also provide a greater functionality than one-to-one communication. In the authors' multicast protocol, the sender directs a message to a named group of receivers, which can be specified by function without requiring the sender to know the specific members of the group. Each host's kernel in the network can respond to every group message sent, providing various levels of reliability. It was found that the overhead of providing dependable multicast over a single local area network was very small, mainly because the protocol operates at the kernel level rather than the user level. Several forms of this multicast communication, expressed as simple message-passing communication primitives, are described, and the effectiveness of the protocol is evaluated using an example of a distributed algorithm. Performance analyses and actual performance data for the protocol are presented.
The use of local area network (LAN) technology for distributed process control is addressed. This paper highlights some current problems, and focuses on some unique new aspects of distributed control which are necessa...
详细信息
The use of local area network (LAN) technology for distributed process control is addressed. This paper highlights some current problems, and focuses on some unique new aspects of distributed control which are necessary for providing a properly integrated approach to distributed supervisory control systems. A subscription service satisfies the need to move data efficiently between supervisory computers, taking into account the possibilities of multiple sources (redundancy) and multiple users of the data, making use of the broadcast capabilities of Ethernet to minimise overhead. A hot-standby pair configuration allows two supervisory computers on the network to operate in a redundant manner, providing high reliability. A global facility secure service allows legislating or locking of access to a shared resource (such as part of a database) on another supervisory computer. This permits multiple operations (which are divisible in time) to be safely done. A virtual display system allows a comprehensive set of operator displays on one supervisory computer to be used by an operator display station on another computer. The concept of servicing the display over the network, rather than transferring the data, results in a reduced load, both for the network and for the computers.
In real-time databasesystems. a transaction may not have enough time to complete. In such cases, partial, or imprecise, results can still be produced. The authors have proposed an imprecise result mechanism for produ...
详细信息
ISBN:
(纸本)0818608757
In real-time databasesystems. a transaction may not have enough time to complete. In such cases, partial, or imprecise, results can still be produced. The authors have proposed an imprecise result mechanism for producing partial results, which is used here to implement timing error recovery in real-time databasesystems. They also present a model of real-time systems.that distinguishes the external data consistency from the internal data consistency maintained by non-real-time systems. Providing a timely response may require sacrificing internal consistency. The authors discuss three examples that have different requirements of data consistency and present algorithms for implementing them.
The authors propose two protocols for transaction processing in quasi-partitioned databases. The protocols are pessimistic in that they permit the execution of update transactions in exactly one partition. The first p...
详细信息
ISBN:
(纸本)0818608757
The authors propose two protocols for transaction processing in quasi-partitioned databases. The protocols are pessimistic in that they permit the execution of update transactions in exactly one partition. The first protocol is defined for a fully partition-replicated database in which every partition contains a copy of every data object. The second protocol is defined for a partially partition-replicated database in which some objects have no copies in some partitions. Both protocols improve their major performance measures linearly with the backup link speed but are not visibly affected by either duration of the partitioning or the database size. This is a desirable property, since the backup link speed is the only controllable parameter.
The authors deal with the task allocation problem in distributedsoftware design, with the goal of maximizing the system reliability. A quantitative problem model, algorithms for optimal and suboptimal solutions, and ...
详细信息
ISBN:
(纸本)0818608757
The authors deal with the task allocation problem in distributedsoftware design, with the goal of maximizing the system reliability. A quantitative problem model, algorithms for optimal and suboptimal solutions, and simulation results are provided and discussed. Because the authors use a new allocation goal--to maximize system reliability--this study complements the existing body of knowledge in task allocation.
The proceeding contains 21 papers. The following topics are dealt with: recovery in distributedsystems.managing replication and network partition;fault-tolerance techniques;fault-tolerant protocols;voting and fault d...
详细信息
ISBN:
(纸本)0818608757
The proceeding contains 21 papers. The following topics are dealt with: recovery in distributedsystems.managing replication and network partition;fault-tolerance techniques;fault-tolerant protocols;voting and fault diagnosis;experimental systems.and, consistency maintenance.
The design and implementation of an experimental fault-tolerant distributeddatabase management system is described. The system provides a logically integrated view of data with distribution transparency and a control...
详细信息
ISBN:
(纸本)0818608757
The design and implementation of an experimental fault-tolerant distributeddatabase management system is described. The system provides a logically integrated view of data with distribution transparency and a controlled data replication. A commitment protocol used to guarantee atomicity of update operations is discussed. Efficient algorithms used to recover a site from a failure and restore data consistency are described. Recovery can be interleaved with the processing of regular database transactions and does not seriously limit the availability of data. The proposed solutions to the problems of fault recovery are designed to take advantage of the properties of a high-bandwidth local area network.
暂无评论