Integrating data sources is a general problem in many scenarios. The main problem is the heterogeneity between data sources which were created and developed separately. In the literature there exist many different approaches to solve that problem. Schema integration approaches derive an integrated schema by resolving conflicts on schema and data model level. Another kind of approaches are multidatabase languages and systems which define and verify views on top of example databases. They try to map local data to a predefined, integrated schema. In this paper we investigate how the schema integration approach GIM and the example-driven approach VIBE basing on FRAQL can be combined for a new combined approach. The combined approach benefits from the powerful mapping language FRAQL and the verification feature by using example databases and from the extensional analysis of the GIM approach. We discuss the advantages of such a combined approach and show its potential in many integration scenarios.
Die Anfrageverarbeitung uber integrierten, heterogenen und verteilten Datenquellen verlangt nach angepassten Vorgehensweisen und Architekturkonzepten. Dieser Beitrag stellt erste Ideen zu einem agentenbasierten Anfrageverarbeitungssystem vor, welches insbesondere auch in der Lage sein soll, in Fehlersituationen (z.B. Ausfall einer Quelle) reagieren zu konnen. Zu diesem Zweck werden Konzepte der adaptiven und dynamischen Anfrageverarbeitung eingesetzt. Die Adaptivitat ist notwendig, damit initial bereits effizient ausfuhrbare Anfragen erzeugt werden. Die Dynamik soll es ermoglichen auch zur Laufzeit einer Anfrage im Fehlerfall noch einzuschreiten, und so die Ausfallsicherheit einer Anfrageausfuhrung zu erhohen. Diese Arbeit diskutiert Moglichkeiten, zur Laufzeit einer Anfrage Datenquellen auszutauschen. Dies ist insbesondere in Situationen sinnvoll, in denen keine Interaktion mit dem Nutzer stattfindet und Anfragen eine lange Laufzeit besitzen (z.B. Publish/Subscribe-Szenario oder die Offline-Verarbeitung von Anfragen mobiler Endgerate). Schlusselworte: dynamische und adaptive Anfrageverarbeitung, Anfrageoptimierung, Agentenarchitektur, virtuelle Datenbanksysteme.
Commercial, non-profit and public organizations are accumulating huge amounts of electronically available text documents. Although composed of unstructured texts, documents contained in archives such as annual reports to shareholders, medical patient records and public announcements often share an inherent, though undocumented structure. In order to enable information integration of text collections with related structured data sources, this inherent structure should be made explicit as detailed as possible. The goal of this study is the establishment of a methodology for the integration of text documents with structured records into a hyper-archive of application-specific entities. The text documents are of implicit structure which has been explicated by data mining techniques as proposed in the DIAsDEM framework for semantic tagging of domain-specific text documents. The result is a probabilistic DTD that serves as a basis for the matching of schemata and for the matching of data instances.
Data Warehousing is not new to Credit Suisse Financial Services. Over the past twenty years, a large number of warehouse-flavored applications was built, ranging from simple data pools to classical management information systems, up to novel customer relationship management applications using state-of-the-art data mining technologies.However, these warehouse projects were neither coordinated nor are they based on the same infrastructure. Moreover, dramatic changes of the business design had a huge impact on information analysis requirements. Both together resulted in a nearly unmanageable complexity.Therefore, Credit Suisse Financial Services started a 3-year enterprise-wide data warehouse re-engineering initiative at the beginning of 1999. This paper presents the motivation, experiences, and open issues of this strategic IT project.
Zusammenfassung Das World Wide Web wurde bisher kaum als Datenquelle f¨ ur Data-Warehouse-Systeme verwendet. In diesem Beitrag wird ein Konzept vorgestellt, das die Potentiale des WWW und des Data Warehouses durch die Einf¨ uhrung standardisierter, im Web abrufbarer Datenw¨ urfel-sogenannte <SUB>"Web Cubes"-zusammenf ¨ uhrt. Die Einf¨ uhrung von Web Cubes erm¨ oglicht den Aufbau eines f¨ oderativen Data-Warehouse-Netzwerkes. Ein solches Netzwerk erlaubt den Austausch von ausgew¨ ahlten Daten zwi-schen Data-Warehouse-Systemen, die Online-Analyse von Webdaten mit Hilfe von OLAP-Technologie und die Integration von Webdaten in ein Data Warehouse. Die gesch¨ aftlichen Anforderungen werden durch Anwendungsf¨ alle und diese unterst¨ utzende Konzepte beschrieben. Die technischen Anforderungen werden daraus abgeleitet und die Verwendung von XML als Metaformat zur Beschreibung von Web Cu-bes diskutiert. Schlie?lich wird der derzeitige Entwicklungsstand von XCube beschrieben. XCube stellt eine Sammlung von XML-basierten Formaten dar, die sowohl Web Cubes selbst als auch Dokumente zur Bearbeitung von Web Cubes spezifizieren. Als Basis f ¨ ur die Modellierung von XCube wurde die Multidimensional Modeling Language (MML) verwendet.
This article presents a case study on the integration of Web s it s based on the HyperView approach. The design of a prototype for a virtual Web site is d e cribed in order to demonstrate the application of theHyperView approach in the field of cultural town information. A number of open issues arising from this particular application are identified and discussed.
Ein Hauptproblem bei der Integration heterogener Datenbestände bilden Konflikte, die durch unterschiedliche Modellierung eines Sachverhaltes der Real-Welt, durch verschiedene Datenmodelle oder auch nur durch unterschiedliche Repräsentation der Real-Welt-Objekte entstehen. Die Konflikte müssen im Rahmen der Integration erkannt und bei der Definition der Abbildung zwischen globalen und lokalen Schemata aufgelöst werden. Da diese Abbildung die Grundlage für die Bearbeitung von Anfragen ist, ergibt sich eine enge Verzahnung von Konfliktbehandlung, Anfragetransformation und -ausführung. Vor diesem Hintergrund wird in diesem Beitrag eine Anfragesprache für Datenbankföderationen vorgestellt und die Behandlung der wichtigsten Konflikte mit den Mitteln dieser Sprache diskutiert.
Mediator-based information systems answer global queries by rewriting them into a combination of queries against physical data sources. One inherent assumption in most systems is that only a combination that satisses the global query completely is considered valid, i.e., it must obtain values for each required attribute. Furthermore, most systems strive for complete answers, i.e., they try to access all relevant sources. These requirements frequently lead to a system behavior that entails a high potential for user frustration: In many scenarios a partially incomplete answer is much more appreciated than no answer at all. Also, obtaining the data from all sources, which can be very costly, is often not necessary. Based on this observation, we developed a cooperative query planning method. For a given query, the set of data sources is selected based on density scores obeying a user-deened price-constraint. The selection process naturally prefers complete answers, but can also cope with incomplete sources. An important result of this study is that, even under the assumption of a very simple cost model, nding an optimal solution is already NP complete. To encounter this complexity, we also include several heuristics. With the development of the Internet and eecient data transfer methods, federated database systems (FDBS) have become a popular means to integrate information from remote information sources SL90]. This FDBS approach aims at querying multiple databases simultaneously through a single, homogeneous interface. For instance, many globally operating companies have set up a federation of their databases to allow queries covering information of all locations. To correctly and satisfyingly answer such queries, the participating databases must be well maintained. Decision makers will not accept query results of a low quality. Incomplete answers will result in incorrect aggregation, missing values may provoke wrong and possibly costly decisions. Recently, the advantages of information integrating systems have attracted a new audience: private information consumers. This new type of users typically neither demands nor expects perfect and complete query results. They rely on simple read-only query interfaces and cooperative behavior of the query engine. Imagine a user searching for a particular book, its authors, publisher, and, if possible, some reviews. Such a user will most likely not nd a missing publisher annoying, and would certainly prefer an answer without a review than no answer at all. This new user proole greatly relaxes the demands on information integration if compared to the typical deployment scenario of federated …
The way online business is performed will be changed by application service providers (ASPs) and ASP marketplaces (MMM). A fundamental shift can be observed in Corporate Application Delivery: The enterprise applications market is transforming into a rental-based software model coupled with managed services [9]. ASP marketplaces provide alternatives to the classical business model of software which will have a significant impact on the software market [10]. Still, before this kind of ASP marketplaces will become commonplace, several critical research questions have to be answered. The main problem is economical, not technical: Application service providers and ASP marketplaces want to know how to set prices for different customer groups and how to offer digital goods in sizeable versions (product lines) which are fitting to the size of the company and the budget of the consumer. Our goal in this paper is to describe different business models for ASP marketplaces such as MMM. We explore ways to version software to product lines. Selling software versions for different prices allows ASP marketplaces to receive the maximum value of digital products from the market, and it leads to inexpensive, disaggregated and customized software packages.
Abstract: During the last few years message queuing and workflow systems have been established as major elements of the software stack. In practice, this middleware is often used to cope with aspects of data federation: Message queuing technology is exploited for application integration. Workflows extend the notion of stored procedures in a federated environment and provide transaction management as well as monitoring features for inter-transaction integrity in such an environment. -Notice: Many of the figures in this paper are take n from [LR] and are copyrighted by Prentice-Hall.
Providing integrated access to multiple information offerings on the Web poses some fundamentally new challenges for practical federated information systems. Other than conventional databases with an explicit and rather stable schema and expressive query APIs, Web-sources deliver semi-structured data without explicit logical structure and semantics, and have rather limited query capabilities. In this paper we present AMetaCar, a federated Web-information system, that provides integrated access to a number of existing Web-catalogues for used cars. At the wrapper level, the sources are continually monitored, and their implicit structure is explicated by means of XML. At the mediator level a hybrid approach is used, which combines a relational DBMS for the regular and common attributes of used-car offers with a persistent XML-DOM (Document Object Model) implementation for the irregular and heterogeneous attributes. XSL is used for presenting XML results. This overall approach combines the scalability of relational databases with the flexibility of XML for creating scalable and maintainable Information Brokering solutions. This work has been partially funded by the ESPRIT-project MIRO-Web (EP 25208).
Kurzfassung: Unter Verwendung der Sprache XML wird ein Ansatz zur Abbildungsbeschreibung erarbeitet, die als Teil einer Integration von Anwendungssystemen die Funktionsabbildung mittels Abhangigkeiten beschreibt. Hierzu werden zunachst die Datenund Funktionsintegration gegenubergestellt und anschliesend eine Integrationsarchitektur vorgestellt. Es wird eine Klassifikation der API als operationale Schnittstelle entworfen, um darauf aufbauend eine generische Methodik zur Abbildungsbeschreibung vorzustellen.
Federated database technology provides a basis to build int egrated engineering environments. The specific requirements of this application dom ain influence the architecture and implementation details of a federated database managem ent system (FDBMS) built to integrate engineering applications. This paper introdu ces an application scenario of an engineering environment, which is used to illustrate the re quirements to be satisfied by an appropriate federated system. New functionality required at the integration level is specified. The influence of the application domain on the following arch itectural aspects is discussed: schema architecture, data model, object materiali sat on and caching, integrity control and update propagation, version control, and the adopt ion of heterogeneous local systems.
In multidatabase systems, a major data integration problem is to resolve data con icts where two objects having the same de nition and representing the same real world object have di erent extensions. Traditional data integration approaches suggest static con ict resolution functions that perform a computation over conicting attribute values, thus assuming a unique and time-independent resolution. In this paper, we argue that such type of data con ict often arises due to heterogeneities among the data capturing and processing techniques and methods used by component databases. That is, component database di er in how and when they map real world data into local data structures. This diversity results in the fact that the quality of the data stored at di erent sites can be di erent and that the quality can also vary over time, thus requiring dynamic data integration methods, depending on which data quality goal is required at the global level. We outline a novel framework that allows to formalize, model and utilize diverse and in particular orthogonal data quality aspects such as timeliness, accuracy, and completeness in database integration. By making the notion of data quality aspects explicit both in modeling and querying a multidatabase system, existing approaches to database integration cannot only be extended, but also tools can developed that ensure di erent types of \high quality data" at the integration level and for global applications.
