A Distributed System is a system with components that although they are linked by a network are geographically scattered. Nevertheless, such a system appears to each of its users to be a single monolithic system both in form and function. The secret of making a dispersed set of components appear to be one single system is, of course, the software that is designed to hold the system together. Ensloe’s model is introduced as a way of expressing the extent to which a system is actually distributed and, so, the size of the task faced by the unifying software in making it appear to be a single system.
The properties required of such a system, such as transparency and scalability are examined, as are the characteristics that are inherent in such a system, notably fault tolerance and resource sharing. Ways of implementing a distributed system so as to ensure that it has these properties are described.
A fully functional Distributed System can be seen as a collection of sharable resources, available to all the users of the system. As such it can be used by a community with access to the supporting network to meet all their information and computational needs. In particular, distributed applications can be constructed especially for their use by combining the relevant resources from those possessed by the system.
●】 Objectives At the end of this module, you should be able • explain what is meant by a Distributed System; • explore the properties required of a Distributed System and examine the ways in which these properties may be ensured; • explain how a network can support not only one Distributed System but also many different Distributed Systems at the same time; • consider the potential for business usage of a fully functional Distributed System.