The use of semantic information in a distributed data structure
Douglas Bruce Bailey · Summit (Simon Fraser University) · 1988
page iii I It is considered that an algorithm which solves a general class of problems will not be as efficient as an algorithm which solves a subset of the class of problems. Similarly, generalized distributed data structures (and their manipulative operations) are less efficient than particular algorithms for distributing data for a specific application. The efficiency limitation of generalized data structures can be partially defeated by using semantic information. This thesis presents a distributed data structure which makes use of information provided by the application to implement an efficient distribution of the data. This information typically includes which nodes in a network are going to produce data, which nodes are going to read data, and how the data is to be distributed and replicated. An implementation of the distributed data structure is given, along with two examples. Analysis of these examples demonstrates the efficiency and flexibility of the data structure.