Showing papers on "Graph database published in 1987"

Thoth-II: hypertext with explicit semantics

Abstract: This paper describes a hypertext system — Thoth-II. This system provides a rich means for modeling semantic interconnections among texts. It allows a user to browse texts, exploring their relations with other texts. These relations are modeled by a directed graph. The texts are embedded in the graph. Connections among specified phrases in the text and the graph structure are automatically formed. In the browsing mode the user is presented with an interactive graphic display of the directed graph. In the text mode the user can use multiple windows to display and interact with the stored text.

Benchmarking simple database operations

Abstract: There are two widely-known benchmarks for database management systems the TP1 benchmarks (Anon et al [1985]), designed to measure transaction throughout, and the Wisconsin benchmarks (Bitton, Dewitt, & Turbyfil [1984]), designed to measure the performance of a relational query processor. In our work with databases on engineering workstations, we found neither of these benchmarks a suitable measure for our applications' needs. Instead, our requirements are for response time for simple queries. We propose benchmark measurements to measure response time, specifically designed for the simple, object-oriented queries that engineering database applications perform. We report results from running this benchmark against some database systems we use ourselves, and provide enough detail for others to reproduce the benchmark measurements on other relational, object-oriented, or specialized database systems. We discuss a number of factors that make an order of magnitude improvement in benchmark performance caching the entire database in main memory, avoiding query optimization overhead, using physical links for prejoins, and using an alternative to the generally-accepted database “server” architecture on distributed networks.

Method and system for retrieval of stored graphs

Abstract: Automation of retrieval of stored graphs in a multi-user system having a central processing facility with processes that create and translate graphical representation into varying levels of readable expression. The created, translated representations are stored in system storage in the form of graph files. The invention is expressed as a method that tabularizes stored graph files by table entries, each identifying a specific graph and including a listing of the corresponding graph files produced for the graph. When a user of the system identifies a graph by table entry and specifies a machine output or modification function to be performed on the identified graph, the tabularized entries are searched to locate an entry for the identified graph. Once the correct table entry is found, the graph file listing is searched according to a user-specified representation preference ordering in the form of a priority schedule to determine whether a specified output function can be performed with any of the stored graph files stored for the identified graph. If a match is found between one of the representations stored for the graph and one of the representations in the priority schedule, the matched file is retrieved from storage and translated to the level required for the function. The translated file is then dispatched to a graphical output device for performance of the function.

Optimizing Join Queries in Distributed Database

Abstract: A reduced cover set of the set of full reducer semijoin programs for an acyclic query graph for a distributed database system is given. An algorithm based on this reduced cover set is then presented which determines the minimum cost full reducer program. We show that the computational complexity of finding the optimal full reducer for a single relation is of the same order as that of finding the optimal full reducer for all the relations. The optimization algorithm is able to handle query graphs where more than one attribute is common between the relations. We also present a method for determining the optimum profitable semijoin program. The computational complexities of finding the optimum cost semijoin program is high. We present a low cost algorithm which determines a near optimal profitable semijoin program. We do this by converting a semijoin program into a partial order graph. This graph also allows us to maximize the concurrent processing of the semijoins. It is shown that the minimum response time is given by the largest cost path of the partial order graph. We can use this reducibility as a post optimizer for the SDD-1 query optimization algorithm. Finally, it is shown that the least upper bound on the length of any profitable semijoin program is N*(N−1) for a query graph of N nodes.

A knowledge-based approach to multiple transaction processing and distributed database design

Abstract: The collective processing of multiple transactions in a database system has recently received renewed attention due to its capability of improving the overall performance of a database system and its applicability to the design of knowledge-based expert systems and extensible database systems. This dissertation consists of two parts. The first part presents a new knowledge-based approach to the problems of processing multiple concurrent queries and distributing replicated data objects for further improvement of the overall system performance. The second part deals with distributed database design, i.e., designing horizontal fragments using a semantic knowledge, and allocating data in a distributed environment. The semantic knowledge on data such as functional dependencies and semantic data integrity constraints are newly exploited for the identification of subset relationships between intermediate results of query executions involving joins, such that the (intermediate) results of queries can be utilized for the efficient processing of other queries. The concept of the conventional query graph is extended to represent distributed transaction executions by the inclusion of site information. A state of the problem space is represented by this extended query graph. The expertise on the collective processing of multiple transactions is embodied into the rules of a rule-based expert system, MTP (Multiple Transaction Processor). These expert rules exploit large amounts of domain-specific semantic knowledge to reformulate the execution plans of queries such that the overall processing cost is substantially reduced. MTP employs the planning technique combined with search method where the plan step infers the necessary constraints, and the search step achieves optimal solutions utilizing the ${\rm A\sp\ast}$ or branch and bound search technique. In the second part, MTP is applied for the determination of horizontal fragments exploiting the semantic knowledge. Heuristics for allocating data in local area networks are developed. In summary, this knowledge-based approach adds to knowledge for the processing of multiple transactions in a distributed environment, and to the design of knowledge-based expert systems which require efficient access to a large knowledge-base implemented on (distributed) relational database systems.

Directed Graph Optical Processor

Abstract: A directed graph processor and several optical realizations of its input symbolic feature vectors and the multi-processor operations required per node are given. This directed graph processor has advantages over tree and other hierarchical processors because of its large number of interconnections and its ability to adaptively add new nodes and restructure the graph. The use of the basic concepts of such a directed graph processor offer significant impact on: associative, symbolic, inference, feature space and correlation-based AI processors, as well as on knowledge base organization and procedural knowledge control of AI processors. Initial iconic alphanumeric data base results presented are most promising.