In the pastfew years, considerable research effort has been spent on data models, processing mo deis, and system architectures for supporting advanced applications Uke CADICAM, software engineering, image processing, and knowledge management. These so-called non-standard applications pose new requirements on database systems. Conventional database systems (i.e. database systems constructed to support businessrelated applications) either cope with the new requirements only in an unsatisfactory way or do not cope with them at ali. Examples ofsuch new requirements are the need of more powerful data models which enable the definition as well as manipulation offairly structured data objects and the requirement of new processing models which better support long-time data manipulation as well as allow database system users to exchange noncommitted results. To better support new data and processing models, new database systems have been proposed and developed which realize object-oriented data models that in turn support the definition and operation of both complex object structures and object behavior. In design environments as the ones represented by CAD applications, these so-called non-standard database systems are usually distributed over server-workstations computer configwations. While actual object versions are kept in the so-called public database on server, designers create new objects as well as new object versions in their private databases which are maintained by the system at the workstations. Besides that, many new design database system prototypes realize a hierarchy of system buffers to accelerate data processing at the system s application level. While the storage subsystem implements the traditional page/segment buffer to reduce the number of I/O-operations between main memory and diste, data objects are processed by application programs at the workstatíon at higher leveis ofabstraction and the objects are kept there by so-called object-oriented buffer managers in special main memory representations. The present dissertation reports on the investigation of database recovery requirements and database recovery performance in design environments. The term design environment is used here to characterize those data processing environments which support so-called design applications (e.g. CADICAM, software engineering). The dissertation begins by analyzing the conanon architectural characteristics of a set of new design database system prototypes. After proposing a reference architecture for those systems, we investigate the properties of a set ofwell known design processing models which can be found in the literature. Relying on both the reference architecture and the characteristics of design processing models, the dissertation presents a thorough study of recovery requirements in the design environment. Then, the possibility ofadapting existing recovery techniques to maintain system reliability in design database systems is investigated. Finally, the dissertation reports on a recovery performance evaluation involving several existing as well as new recovery mechanisms. The simulation model used in the performance analysis is described and the simulation results are presented.
| Identifer | oai:union.ndltd.org:IBICT/oai:lume56.ufrgs.br:10183/115971 |
| Date | January 1989 |
| Creators | Iochpe, Cirano |
| Contributors | Universitaet Karlsruhe. Fakultaet Fuer Informatik, Lockemann, Peter |
| Source Sets | IBICT Brazilian ETDs |
| Language | English |
| Detected Language | English |
| Type | info:eu-repo/semantics/publishedVersion, info:eu-repo/semantics/doctoralThesis |
| Format | application/pdf |
| Source | reponame:Biblioteca Digital de Teses e Dissertações da UFRGS, instname:Universidade Federal do Rio Grande do Sul, instacron:UFRGS |
| Rights | info:eu-repo/semantics/openAccess |
Page generated in 0.0024 seconds