The morphology and architecture of a distributed data system
T. A. King, Steven P. Joy, Raymond J. Walker · AIP conference proceedings · 1993
Members of the Planetary Plasma Interactions Node of the Planetary Data System (PDS/PPI) have designed and implemented a data system which allows for distributed access to centralized data, centralized access to distributed data, and distributed access to distributed data. While these types of data access might seem like completely different problems, they are not. The only differences are the number of access pathways to and from the data. By adopting a client/server model of system design we have been able to provide transparent access to both distributed and centralized data. In addition, by carefully choosing the location of the division between the client and server, the response of the system as a whole can be greatly enhanced. For example, the user interface is a stand‐alone application which configures itself according to information stored with the data sets and the data sets may be local or distributed. All keystroke interactions are handled on the local machine and the end results of menu selections are transferred to and from the client and server. This model differs from those used by other distributed systems, but in the case of interactive networked applications it proves the most efficient. Also discussed in this paper is how data from multiple sources can be brought together and delivered as a single unit, how large scale data bases can easily be handled, and how data maintained in a variety of database management systems can be transparently accessed.