|
|
FDBMS Research Group Publications |
In Symposium on Objects and Databases in 14th European Conference on Object-Oriented Programming (ECOOP 2000). Sophia Antipolis and Cannes (France), June 2000. (© Springer-Verlag, Lecture Notes in Computer Science, Vol. 1813, Springer, 2000)
In this paper, we try to outline the goodness of using an O-O model on designing multidimensional Data Marts. We argue that multidimensional modeling is lacking in semantics, which can be obtained by using the O-O paradigm. Some benefits that could be obtained by doing this are classified in six O-O-Dimensions (i.e. Classification/Instantiation, Generalization/Specialization, Aggregation/Decomposition, Caller/Called, Derivability, and Dynamicity), and exemplified with specific cases.
In Int. Workshop on Engineering Federated Information Systems (EFIS 2000). Dublin (Ireland), June 2000.
This paper is devoted to Data Warehousing architecture and its data schemas. We relate a federated databases architecture to Data Warehouse schemas, which allows us to provide better understanding to the characteristics of every schema, as well as the way they should be defined. Because of the confidentiality of data used to make decisions, and the federated architecture used, we also pay attention to data protection.
Technical Report LSI-00-26-R. Dept Llenguatges i Sistemes Informàtics (Universitat Politècnica de Catalunya), April 2000.
Extended version of the previous paper: "Information System Architecture for Data Warehousing from a Federation".
Novatica. ATI, 1999 (In Spanish)
En este trabajo se presenta nuestra propuesta de creación de un prototipo de sistema cooperativo para la integración de fuentes heterogéneas de información y almacenes de datos en el cual se centran actualmente nuestras investigaciones. El objetivo general es proporcionar una capa de software que permita la cooperación entre diversas fuentes de información que están interconectadas mediante una red de líneas de comunicación. Cada fuente posee sus propios servicios de respuesta a preguntas que sobre sus datos realizan sus usuarios y, adicionalmente, se desea ofrecer a determinados usuarios la capacidad de acceder al conjunto de datos de una forma uniforme (acceso integrado), ya sea en tiempo real, ya sea a través de almacenes de datos.
Report of EFIS'99 Workshop". ACM SIGMOD Record, 28 (3): pp 9-11 (1999).
Novatica #140 (special isue on Bases de Datos Avanzadas), pp 5-6. ATI, Julio-Agosto 1999 (in Spanish).
Report LSI-99-34-R, Dept LSI, UPC, Barcelona 1999.
The BLOOM (BarceLona Object Oriented Model) data model was developed to be the Canonical Data Model (CDM) of a Federated Database Management System prototype. Its design satisfies the features that a data model should have to be suitable as a CDM. The initial version of the model (BLOOM91) has evolved into the present version, BLOOM99.
This report specifies the syntax of the schema definition language of BLOOM99. In our model, a schema is a set of classes, related through two dimensions: the generalization/specialization dimension, and the aggregation/decomposition dimension. BLOOM supports several features in each of these dimensions, through their corresponding metaclasses.
Even if users are supposed to define and modify schemas in an interactive way, using a Graphical User Interface, a linear schema definition language is clearly needed. Syntax diagrams are used in this report to specify the language; an alternative using grammar productions appears as Appendix A. A possible graphical notation is given in Appendix B.
A comprehensive running example illustrates the model, the language and its syntax, and the graphical notation.
In: ECOOP Workshop (Proc. ECOOP Workshops & posters, ECOOP'99, Lisbon, June 1999).
Once argued the desirable characteristics of a suitable CDM, the BLOOM model (BarceLona Object Oriented Model) was progressively defined. It results in an extension of an object oriented model with a semantically rich set of abstractions. BLOOM was not developed as a whole but suffered extensions in different phases. Its abstractions were conceived for building the FDBS in as needed basis. It drove to a lack of unity and differences in the nomenclature.
The necessity of revising the BLOOM model outcropped during the design process of the directory of the FDBS. It is essential to have such storage system because of the amount of needed information in building and operating a FDBS. The directory is the core of our FDBS architecture and it must contain the different schema levels as well as the mappings among them. Therefore, the model had to be fixed in order to store those schemas and mappings in a structured manner.
In: Conrad, Hasselbring & Saake (eds.): Engineering Federated Information Systems (Proc. 2nd Int. Workshop EFIS'99, Kuehlungsborn, May 1999). infix, Sankt Augustin, 1999, pp 1-4 (ISBN 3-89601-013-1)
In: Proceedings, 9th Int. Conf. on Database and Expert Systems Applications (DEXA98, Viena, August 1998), pp 746-756.
In: B. Eaglestone, B. Desai & J. Shao (eds.): Proceedings of 1998 Int. Database Engineering and Application Symposium (IDEAS98, Cardif, July 1998). IEEE-CS Press, pp 150-159.
External schemas provide logical data independence in OODBs. External schemas may contain conceptual schema classes as well as derived classes defined either directly o indirectly from conceptual schema classes. The classes upon which a derived class is defined are its base classes. A derived class may be defined either by object preserving semantics if it only contains objects of its base classes; or by object generating semantics if it contains new objects generated from the objects of its base classes. Defining derived classes by object generating semantics makes it possible to carry out sophisticated reorganisations of existing information which would otherwise be impossible -i.e. transformation of values into objects, or aggregation of objects to form a new conceept. In this paper will shall study two of the main problems of defining derived classes: the generation of identifiers for the objects of the derived classes; and the transmission of modifications between the objects of the derived classes and those of the base classes. In a schema evolution environment we consider it necessary to be able to define partially derived classes: derived classes which migth also contain non-derived elements; this possibility presents additional problems which are dealt with in an integrated way.
In: T. Ozsu, A. Dogac & O. Ulusoy (eds.): Issues and Applications of Database Tecnology (Proc 3rd. Conference on Integrated Dessign and Process Tecnology, IADT/IDPT, Berlin, July 1998. IDPT-Vol 2, 1998, Society for Dessign and Process Science, Austin, TX, pp 272-279.
Proceedings, 6th Int. Conf. on Extending Database Technology, Valencia, March 1998. Springer, LNCS 1377, 1998.
In: R. Wagner (ed.): Proceedings, DEXA Workshop on Database and Expert Systems Applications (DEXA'97), Toulouse, September 1997. IEEE Computer Society Press, 1997, pp 516-522.
External schemas are derived from the database conceptual schema; they can be used to simulate changes in it. Sometimes the final users' information requirements change: they need new information which cannot be derived from the information in the database. The solution put forward here is the definition of partially derived classes (capacity augmenting classes) which may contain non-derived information in the intension and moreover unlike other systems in the extension. When an external schema with partially derived classes is to be defined, the conceptual schema has to be modified in order to include to include the non-derived information. In order to avoid unnecessary modifications of the conceptual schema the use of a test environment for the definition of temporal external schemas is also proposed.
In F. Baader, M.A. Jeusfeld & W. Nutt (eds), Proceedings of the 4th Workshop on Knowledge Representation meets DataBases (KRDB-97), Athens, Greece, August 30, 1997, pp 15:1-15:7, Volume CEUR-WS/Vol-8, 1997.
The process to provide integrated access to several, independent information sources is not easy, due to semantic heterogeneities which lead to semantic conflicts or contradictions. In this paper we present a discussion about the existence of contradictions, their importance and how they can be handled when the need of cooperation arises.
Invited talk (in Spanish), actas 2as Jornadas de Investigacion y Docencia en Bases de Datos (JIDBD97, Getafe/Madrid, July 1997). Universidad Carlos III de Madrid, p. 7.
ACM SIGMOD Record 26(4), pp 54-56 (1997).
In S. Conrad et al. (eds), Proceedings of the International CAiSE'97 Workshop on Engineering Federated Database Systems (EFDBS'97), pp 93-104. Otto-von-Guericke-Universitat Magdeburg, Fakultat fur Informatik, 1997.
This paper presents an approach for thigtly federated database systems with several federated schemas that consist of a schema architecture and a functional architecture. The reference schema architecture has been extended to deal with security aspects not well solved previously. The functional architecture includes the processors needed to build the federation and the main modules of the execution architecture.
In: 10th ERCIM Database Research Group Workshop on Heterogeneous Information Management, Prague 1996, ERCIM-96-W003 ERCIM, 1996, pp 43-49.
Invited chapter in: José Cuena(ed), Panorama Informático. Federación Española de sociedades de Informática FESI, 1996, pp 39-64.
Proceedings of the 1996 International Conference on Object Oriented Information Systems (OOIS'96), London, December, 1996, pp 317-332.
In this paper two external schema generation algorithms are presented. These algorithms are defined in the framework of the new external schema definition methodology for OODBs put forward in (Samos, 1995): they generate an external schema from a set of classes selected from those existing in a external schema definition repository. Their main characteristic is that they avoid the generation of auxiliary classes that are not included in the defined external schema.
Submitted for publication.
This paper explores the possibilities of recovering object behavior from non object oriented (OO) applications. The objective of this paper is to survey current research on object recovery, with an special emphasis aon approaches that contemplate the behavior of the object. A framework for evaluation and comparison is constructed and applied to ten proposals. The framework includes several descriptive criteria and the activities usually performed during the object behavior recovery process.
In Yetongnon and Hairi (eds): Parallel and Distributed Computing Systems (Proc. PDCS'96, Dijon). ISCA, 1996, pp 766-771.
Different schema architectures for federated database systems have been proposed, particularly the 5-level reference architecture by Seth & Larson. We analyze three situations in which this architecture has shortcomings: databases components of several federations, external schemata in user models different from the canonical model, and multiple semantics at the federated schema level, and we provide better solutions to each of them. With these solutions a new 8-level schema architecture framework is developed.
Proceedings 7th. Int. DEXA Workshop on Database and Expert Systems Applications , Zurich, September 9-10, 1996, pp 313-318.
In this paper the definition of deductive conceptual models (DCMs) using Prolog in order of specify different aspects of OODBs is proposed. The result of the specification process using this technique is an executable prototype of the system. Having a prototype directly available, along with the system specifications, is particulary useful in order to define additional elements in the context of OODBs (e.g. schema evolution, definition of derived classes, definition of external schemas). The use of this technique is proposed mainly due to the difficulty of building prototypes of the mentioned elements over commercial OODBs. A brief example of a practical application of this technique -the specification of a conceptual schema definition system and its associated data model- is presented.
Actas de las Primeras Jornadas en Investigacion y Docencia en Bases de Datos A Coruña, 1996, pp 272-282 (in Spanish).
La creciente necesidad de cooperacion entre multiples bases de datos heterogeneas y autonomas, requiere proveer acceso integrado a sus datos como si de una sola base de datos se tratase, formando un sistema de bases de datos federado. En este articulo mostramos la problematica del acceso integrado y el trabajo realizado por nuestro grupo de investigacion en el area de sistemas de gestion de bases de datos federadas fuertemente acopladas con multiples federaciones.
Proceedings 2nd. Int. Conf. on Object Oriented Information Systems (OOIS'95, Dublin), 1995, pp. 154-166.
In this work, a new external schema definition methodology for object oriented databases (OODB) that solves some of the problems of existing methodologies is presented. In particular, this new methodology avoids the need for generating derived classes that are not directly used in the defined external schema. Also, it offers a way of integrating and reusing different external schema definition.
Object Oriented Databases (Proc. 7th EDRG Workshop, Lisbon, 1995). Report ERCIM-95-W001, ERCIM, Le Chesnay, 1995, pp 40-43.
S. Laufmann, S. Spaccapietra and T. Yokoi (eds). Proc. of the 3rd International Conference on Cooperative Information Systems (CoopIS'96,Vienna), Univ.Toronto , 1995, pp. 19-31.
T. Cheung (ed.): Proc. 6th Int. HK DB Workshop: Database Reengineering and Interoperability (Hong Kong, 1995), City University of Hong Kong, 1995, pp 243-254.
Proc. 11th Int. Conference on Data Engineering (ICDE'95, Taipei). IEEE-CS Press, 1995, pp 505-512.
The process of integrating the schemas of several databases into an integrated schema is not easy, due to semantic heterogeneities. We present a method to detect class similarities by following a strategy and applying comparison criteria, that exploits the semantically rich structures of the schemas (previously enriched), along both the generalization/specialization and the aggregation dimensions. Relaxations may be applied to conform a pair of classes , resulting in penalizations in the computation of the degree of similarity. Our approach needs less comparisons than methods based on attribute comparison.
Invited chapter in: Bukhres & Elmagarmid (eds.) Object Oriented Multidatabases . Prentice-Hall, 1995.
Proceedings of the Third International Conference on Parallel and Distributed Information Systems (PDIS'94, Austin, 1994). IEEE-CS Press, 1994, pp 213-216.
D. Karagiannis (ed.): Database and Expert Systems Applications (5th International Conference DEXA'94, Athens, 1994). Springer Verlag, LNCS 856, 1994, pp 125-134.
A method to convert relational schemas to an object oriented semantic model is presented. It makes explicit knowledge about the semantic of the database that was implicit in the schemas or in the extenxion, through a knowledge adquisition phase. The method is general, semiautomatic, and has a number of applications.
H. Jaakkola (ed.) Information Modelling and Knowledge Bases VI (4th European-Japanese Seminar on Information Modelling and Knowledge Bases, Kista, June 1994). IOS Press, Amsterdam, 1995, pp 447-467.
The capabilities of a data model to represent any conceptualization (expressiveness) are very important for faithful modelling. Focusing on the generalization/specialization dimension, unique features of the BLOOM model are presented: specialization criteria, four kinds of specialization, usage of specialization metaclasses, transitive specializations, combinations of specializations that automatically generate specialization semilattices. These features are exemplified through a complete example.
UPC, Report LSI-93-19-T. Barcelona, 1993 (in Spanish). Replaced by #28.
Deductive and Interoperable Databases (Proc. 6th EDRG Workshop, Barcelona, 1994). Report ERCIM-94-W005,ERCIM, Le Chesnay, 1994, pp 8-25.
Proceedings of the 14th International Conference on Information Systems (ICIS'93, Orlando 1993), pp. 247-254.
Diversity is an asset, as long as it allows cooperation. In the case of information systems and databases, the data model used os a help or a hindrance for this cooperation of diverse views; this is characterized by the semantic relativism of the model. We first analyze diversity within an information system, where cooperation is made possible by the use of external schemata; semantic relativism of the model of the database schema is the key factor. Then we discuss diversity between different information systems, where they cooperate thru interoperability, by schema integration into federated schemata; semantic relativism of the canonical data model is shown to be determinant.
Jaakkola, Kangassalo, Kitahashi & Markus (eds) Information Modelling and Knowledge Bases V (3rd European-Japanese Seminar on Information Modelling and Knowledge Bases, Budapest, 1993). IOS Press, Amsterdam, 1994, pp. 400-420. (Also in: U.P.C., Report LSI-93-2-R).
Relational databases schemas must be semantically enriched to reflect knowledge about the data, as needed by many applications. One technique is the analysis of the database extension, extracting existing data dependencies. In this paper, new algorithms to extract functional and inclusion dependencies are presented; they are designed to minimize the number of disk accesses.
Worboys & Grundy (eds) Advances in Databases (Proceedings, 11th British National Conference on Databases (BNCOD-11), Keele, 1993). LNCS 696, Springer, 1993, pp. 58-75.
Integrated access to a Federated Database System, whatever its architecture is, requires a deep knowledge about the semantics of its component databases so that interdatabase semantic relationships can be detected. Unfortunately, very often there is a lack of such a knowledge and the local schemas, being semantically poor as a consequence of the limited expressiveness of traditional data models, do not help to adquire it. The solution to overcome this limitation is to upgrade the semantic level of the local schemas through a semantic enrichment process that discovers implicit knowledge and makes it explicit by converting the local schemas to rich schemas expressed in a canonical model. Here we a methodology for semantic enrichment consisting of two phases. In the knowledge acquision phase, restrictions in the form of different kinds of identifiers and dependencies are discovered by analyzing the intension and the extension of the database. Then, in the conversion phase, the schemas augmented with this knowledge are converted to rich schemas expressed in a canonical object oriented model.
Schek, Sheth & Czejdo (eds) Proceedings, 3rd Int. Workshop on Research Issues in Data Engineering: Interoperability in Multidatabase Systems (RIDE IMS-93, Vienna, 1993). IEEE-CS Press, 1993, pp.126-129. Expanded version in "A Methodology for Semantically Enriching Interoperable Databases".
Integrated access to a Federated Database System, requires a deep knowledge about the semantics of its component databases so that interdatabase semantic relationships can be detected. Unfortunately, very often there is a lack of such a knowledge and the local schemas do not help to adquire it. The solution is upgrade the semantic level of the local schemas thru a semantic enrichment process where implicit knowledge is discovered and made explicit. Here we outline a methodology for semantic enrichment consisting of two phases. In the knowledge acquision phase, restrictions in the form of different types of identifiers and dependencies of several kinds are discovered by analyzing the intension and the extension of the database. Then, in the conversion phase, the schemas augmented with this knowledge are converted to rich schemas expressed in a canonical object oriented model.
Sheth & Czejdo (eds) Proceedings, 3rd Int. Workshop on Research Issues in Data Engineering: Interoperability in Multidatabase Systems (RIDE IMS-93, Vienna, 1993). IEEE-CS Press, 1993, pp.26-33.
The integration of database schemas into a federated one involves a detection phase where the similarities that exist among the classes of the different databases must be discovered in order to determine their semantic relationships. This phase is typically characterized by the complexity of the comparison task and the deep knowledge about the semantics of the databases that is required to perform it. here we present our approach to handle this problem bu upgrading the semantic level of the local schemas and guiding the search of the comparison process by the structure of the generalization/specialization semilattices and aggregation graphs of the resulting rich schemas. Unnecessary as well as most promising comparisons are identified and a systematic procedure to analyze the rich component schemas in a meaningful way is given. A reduction of the complexity is achieved. The methodology serves as the basis for a tool that semiautomatizes the process. The expressivenes of our canonical model plays a central role.
Hsiao, Neuhold & Sacks-Davis (eds) Interoperable Database Systems (DS-5) (Proceedings, IFIP WG2.6 Database Semantics Conf. on Interoperable Database Systems (DS-5), Lorne, Victoria, Australia, 1992). North-Holland, 1993, pp.191-205.
An important kind of semantic conflicts arising in database interoperability are schematic discrepancies, where data in one database correspond to a metadata in another. We present a general solution to these discrepancies, based on a framework of two dimensions: generalization and aggregation. Operations to transform metadata into data and vice versa are defined, in the relational model and in a object-oriented model. These operations may be applied at different levels in a federated architecture.
Ozsu, Dayal & Valduriez (eds) Distributed Object Management (Proceedings, Int. Workshop on Distributed Object Management, IWDOM, Edmonton, Canada, 1992). Morgan Kaufmann 1994, pp.309-314.
In the operability context, autonomous pre-existing databases are federated at some degree in order to share their data. the growing importance that the operability among object-oriented and relational databases is gaining, makes essential the development of adequate data models that serve as canonical models for the federation.. The canonical model constitutes the basic building block for methodologies and tools for the detection of the interdatabase semantic relationship that is required to access the federation in a integrated way. This paper presents one such canonical model and points out its usage for coupling relational and object-oriented databases thru a semantic enrichment process.
Solutions to Problems of Scientific Databases (Proceedings, 2nd EDRG Workshop, Abingdon, England, 1992). ERCIM, RAL, Abingdon, 1992. Expanded version in #12.
Encuentro sobre Bases de Datos en la Administracion Publica. Ministerio para las Administracions Publicas, Madrid, 1992, pags 195-200 (In Spanish).
ACM SIGMOD Record vol 20, #4, pp. 44-48 (special refereed issue: A. Sheth (guest ed.): Semantic Issues in Multidatabase Systems, Dec. 1991).
We develop a framework of characteristics, essential and recommended, that a data model should have to be suitable as canonical model for federated databases. This framework is based on the two factors of the representation ability of a model: expressiveness and semantic relativism. Several data models are analyzed with respect to the characteristics of the framework, to evaluate their adequacy as canonical models.
U.P.C., RR LSI-91-22. Barcelona, 1991.
Novatica vol XVII #91 (Special issue Bases de datos avanzadas). ATI Barcelona, 1991 (In Spanish). Replaced by #9.
Proceedings 1st Int. Workshop on Interoperability in Multidatabase Systems (IMS-91, Kyoto, 1991). IEEE-CS Press, pp.222-225.
Federation Rules to be followed when associating external schemas into federated schemas are formulated. Discriminated operations in an object oriented model, such as discriminated generalization, and in the relational model, such as discriminated union and outer union, are introduced. It is then shown how they may help in obeying the Federation Rules. These operations allow multiple semantics in a single federated schema, and thence minimize the number of federated schema needed, as well as their maintenance. They also allow easier solutions to other problems.
Proceedings 1st Int Workshop on Interoperability in Multidatabase Systems (IMS-91, Kyoto, 1991). IEEE-CS Press, pp. 71-78.
When a database is to be interoperated with others, its schema must first be converted to the canonical model of the Federated system. This implies a knowledge acquisition process to upgrade the semantic level of the schema. This paper presents a methodology to convert relational schemas to a semantically rich object oriented model. It is based on inclusion dependencies, but also takes into account exclusion and complementariness dependencies.
U.P.C., RR LSI 89.18. Barcelona, 1989 (In Spanish).
Univ. Politecnica de Catalunya, FIB RR 18.87. Barcelona, 1987.
Proceedings of the 2nd Int. Conf. on Data Engineering (Los Angeles, 1986). IEEE Press.