Norman Revell

An Algebra for a Temporal Object Data Model

In this paper, we present a temporal object data model, which has been adapted from the unified model of OODB and RDB in UniSQL/X so that a time dimension can be easily added to form temporal relational-like cubes but with aggregation and inheritance hierarchies. A query algebra, that accesses objects through the associations of aggregation, inheritance and time-reference, is thereby defined. Due to the adaptation of the unified model of RDB and OODB, the temporal object data model supports both homogeneity and heterogeneity in the time dimension, and the algebra reflects the spirit of both temporal relational algebra and object algebra. Data query examples through “The Wood Panel Deformation Measurement Database” illustrate algebraic operations and a brief evaluation of algebra has been given.

FIBEX, an extractor enabling querying of documents using SQL

With the increasing use of digital libraries, querying of file systems has seen renewed interest, as file systems are commonly used as repositories of documents for digital libraries. We aim to extend DBMS querying capabilities to file systems by employing a data extractor called FIBEX (File Base Extractor). FIBEX generates index files from previously semi structured document files. Browsing techniques and behaviours observed in a Digital Library will be incorporated into the process of generating the FIBEX index files, to improve query efficiency. The paper describes the influence of the browsing techniques of digital library users on the creation of the index files by FIBEX.

Database Performance Evaluation: A Methodological Approach

This paper presents the CITY benchmark. Its design is based on the results of several in-depth studies at several large UK organisations. These studies examined over 4800 applications, and investigated over 40,000,000 transactions accessing around 5000 tables. The work revealed a common access pattern between all the organisations, despite the different nature of these environments. The benchmark shows significant deviation from the established TPC-A script and the TPC-B script when tested in a range of environments, ranging from standalone PC to large mainframe computers and multiprocessors.

A Measure for Measure

Comparing the behaviour of the workloads of different transactions posed a problem for this research because the review of literature could not find a well-accepted comparison technique that could be relied on for such comparisons. This paper proposes the use of the rate of change of DBMS behaviour against a test parameter as a common basis of comparison between different transaction workloads. The hypothesis is based on the relationship between the rate of change of a specific workload and the degree of load it applies on the tested parameters, which implies less test time and cost.

Facts and Myths of Browsing and Searching in a Digital Library

In recent times, there has been increased interest in the querying of digital libraries (DLs). This is due in part to the development of the WWW, which enables easy access to both centralised and distributed digital library sources. The majority of published works on querying DLs are associated with information retrieval (IR), also known as digital querying. Information retrieval techniques are popular with querying DLs due to their flexibility in querying semi-structured data. In contrast, database querying of DLs has been largely ignored until only recent years. The key aspects of database querying of DLs involve the integration of database querying with browsing or navigating techniques to query semi-structured data. Our interest lies in developing the relatively limited database query facilities currently available to users of DLs, and a key stage in this process is to define what kinds of searching and browsing typical users would like to perform.

A uniform framework for processing temporal object queries

The paper explores a uniform framework for processing temporal queries in the context of object oriented databases. A temporal object data model is developed by extending the unified model of RDBs and OODBs from UniSQL/X with a time dimension, that forms temporal relational-like cubes but with aggregation and inheritance hierarchies. A query algebra is thereby defined to provide an access of objects through these associations of aggregation, inheritance and time reference. Due to the hierarchy of the data model and reducibility of the query algebra, an extensible approach to processing temporal object queries within the uniform query processing framework is attained. A set of transformation rules is identified for query rewrite, that includes the known relational and object rules plus those pertaining to the time reference. To evaluate temporal queries involved in the path, a decomposition strategy is proposed. It has been shown that temporal object queries can be processed and optimized within the uniform query processing framework.

Query processing and optimization in temporal object-oriented databases

Investigates temporal query processing and optimization in the context of object-oriented databases. Based on our temporal object data model and algebra, a strategy of decomposition is presented to process temporal queries that involve associations of both aggregation hierarchy and time-reference. That is, the evaluation of an enhanced path, which is defined to extend a path with time-reference, is decomposed by initially dividing the path into two sub-paths: one contains the time-stamped class that can be optimized by making use of the ordering information of temporal data and another in an ordinary sub-path which can be further decomposed and evaluated using different algorithms. The intermediate results of traversed two sub-paths are then joined together to create the query output. Algorithms of time-related operations and four basic join algorithms have been implemented with stream processing techniques and presented with a cost analysis. Optimization issues have been discussed.

Processing temporal queries in the context of object-oriented databases

Abstract This article investigates an extensible approach to processing temporal queries in the context of object-oriented databases. Within the uniform query processing framework, a strategy of decomposition is proposed for processing temporal queries that involve paths based on the defined temporal object data model and query algebra. Algorithms for processing the decomposed query components have been implemented using stream processing techniques and are presented with cost analyses. Heuristics that optimize the temporal queries are also presented. Both cost analysis and simulations show that join time cost is linearly increased with the expansion in the number of time-epochs and that utilising the heuristics presented in this article can lead to a significant time cost saving.