摘要:

We describe a new kind of distributed system service, the availability management service, responsible for ensuring that the critical services of a distributed system remain continuously available to users despite arbitrary numbers of concurrent node removals and node restarts caused by failures, maintenance, and growth. We stress the main ideas behind this new service, and outline a simple design that depends on the existence of synchronous membership and atomic broadcast group communication services. Extensions of this initial design to deal with asynchronous group communication services are also briefly discussed.

DOI:10.1049/sej.1993.0009

出版商:IEE    

年代:1993

数据来源: IET

摘要:

In this paper, we take a reuse-oriented perspective to designing configurable distributed applications. An application domain is defined as a family of systems with some features in common and others that differentiate them. During domain modelling, reusable specifications and architectures are developed, which capture the similarities and variations of the family of systems that compose the application domain. Target systems are generated by tailoring the reusable specifications and architectures, given the requirements of the target system. In this paper, we describe how the domain modelling approach to developing reusable software may be applied to the development of configurable distributed applications. We also describe criteria for distributed subsystem structuring and for distributed subsystem configuration. An example is provided of applying this approach.

DOI:10.1049/sej.1993.0010

出版商:IEE    

年代:1993

数据来源: IET

摘要:

Darwin is a configuration language which allows distributed and parallel programs to be structured in terms of groups of process instances which communicate by message-passing. In addition to expressing static structure, Darwin can be used to express structures which change dynamically as execution progresses. In this paper, we present a set of examples illustrating the use of Darwin in constructing parallel programs. As processes can be considered as abstractions of physical processors, Darwin can also be used to describe the hardware structure of distributed memory multicomputers in terms of processors and their interconnection. We illustrate this for a multicomputer constructed from transputers and show its use in the process of mapping the logical structure of a parallel program to the physical hardware.

DOI:10.1049/sej.1993.0011

出版商:IEE    

年代:1993

数据来源: IET

摘要:

Durra is a language and runtime support system for developing distributed applications. In this paper, we describe the general nature of these applications, and the need for separating structural and behavioural details. A Durra user describes an application as a set of components (application tasks and communication channels), a set of alternative configurations showing how these components are connected at runtime, and a set of conditional configuration transitions that take place at runtime. We illustrate these features through examples drawn from the fault tolerance domain. We conclude with an illustration of the use of Durra in an incremental software development scenario.

DOI:10.1049/sej.1993.0012

出版商:IEE    

年代:1993

数据来源: IET

摘要:

Surgeon is a tool that helps programmers prepare software to be dynamically reconfigured.Dynamic reconfigurationenvironments allow programmers to change a distributed software application while it executes, but existing environments require programmers to manually adapt their software for reconfiguration. However, recent work on the automatic preparation of software for a distributed execution environment (software packaging) shows that interfacing software can be generated automatically, given abstract descriptions of the components and the desired application geometry. Surgeon merges these results so that programmers can dynamically reconfigure their applications without the overhead of manually adapting the software.

DOI:10.1049/sej.1993.0013

出版商:IEE    

年代:1993

数据来源: IET

摘要:

Early work on reconfiguration of distributed applications was carried out as part of the Argus project at MIT. In this paper, we review that work, focusing on correctness conditions and on alternative implementations, and discuss lessons learnt, not only about the reconfiguration system, but also about language design for supporting reconfigurable systems.

DOI:10.1049/sej.1993.0014

出版商:IEE    

年代:1993

数据来源: IET