The concept of serializability has been the traditionally accepted notion of correctness in database systems. However, in a heterogeneous distributed database system (HDBMS) environment, ensuring serializability is a difficult task mainly due to the desire of preserving the local autonomy of the participating local database systems. In this paper, we introduce a new correctness criterion for HDBMSs, two-level serializability (2LSR), and show that it preserves database consistency in several HDBMS models. Further, we present a simple protocol for ensuring that schedules in an HDBMS environment are 2LSR. This protocol is easily implementable and does not violate the local autonomy of sites. 1 Introduction The problem of transaction management in a heterogeneous distributed database management system (HDBMS) has received considerable attention from the database community in recent years. The basic problem is to integrate a number of pre-existing local database management systems (DBMSs)...

A method for integrated concurrency control and recovery, applicable to heterogeneous multidatabase systems is proposed 1 . The role of the participant in the two-phase commit protocol is laid on an entity called 2PC Agent associated with the local database system. The main importance of the method is in preserving global serializability in the presence of unilateral aborts and site failures. The method requires the participating database systems to use the strict two-phase locking or a comparable rigorous concurrency control policy. Introduction There has recently been much interest in integrating pre-existing databases managed by heterogeneous database management systems (DBMS). This is understood to be caused by the need to eliminate "islands of information" [11] and, generally, the necessity to improve the interoperability of database systems [23]. There are various architectures supporting these objectives. The multidatabase architecture [25] is characterized by preserving vari...

The integration of information from multiple databases has been an enduring subject of research for almost 20 years, and many di erent solutions have been attempted or proposed. Missing from this research has been a uniform framework. Usually, each solution develops its own ad-hoc framework, designed to address the particular aspects of the problem that are being attacked and the particular methodology that is being used. To address this situation, in this paper we de ne a formal model for multidatabases, which we call Multiplex. Multiplex is a simple extension of the relational model, which may serve as a uniform abstraction for many previous ad-hoc solutions. Multiplex is based on formal assumptions of integrability, which distinguish between scheme and instance reconcilability among independent databases. Multiplex supports database heterogeneity, and it provides several degrees of freedom that allow it to model actual situations encountered in multidatabase applications. In addition, in situations in which a single answer is not obtainable (either because the global query is not answerable, or there are multiple candidate answers), Multiplex de nes approximative answers. Finally, Multiplex provides a practical platform for implementation. A prototype of such an implementation is described brie y.

Abstract not found

Algorithms for scheduling of distributed transactions in a heterogeneous multidatabase, in the presence of failures, are presented. The algorithms of prepare certification and commit certification protect against serialization errors called global view distortions and local view distortions. View serializable overall histories are guaranteed in the presence of most typical failures. The assumptions are, among others, that the participating database systems produce rigorous histories, e.g. by using the strict two-phase locking policy, and that no local transaction may update the data accessed by a global transaction that is in the prepared state. The main advantage of the method, as compared to other known solutions, is that it is totally decentralized. 1 Introduction A multidatabase system can be seen as a system achieving some degree of interoperability of pre-existing local database systems (LDBS). For an overview of issues in multidatabases see e.g. [10] and [7]. In the following, ...

Negotiation and contracting have been traditionally used in the database context in connection with passing schema information between interconnected database nodes. Several protocols have been proposed to address the transmission of schema information. However, all of the approaches suggested so far are fairly simplistic, restrictive in nature as they do not result in alleviating conflicts or formulating partial solutions in a collective manner. In contrast, Distributed AI (DAI) has taken the view that negotiation is a mechanism used by autonomous systems to resolve inconsistent views and reach agreement on how they can work together in order to cooperate effectively. Although this approach may sound appealing it becomes an intractable problem in an environment as complex as interoperable database systems. In this paper we reassess the concepts of negotiation and contracting in both distributed databases and DAI and propose a flexible framework tailored around the client/server model ...

Many organizations are working to deploy cooperative information systems (CIS). Typically a single entity controls all of the cooperating systems. Recently, however, more attention has been focused on systems that require a new level of coordination and cooperation --- systems which are not wholly controlled by any single entity. We describe one particular case --- an information marketplace. consisting of multiple, independentlyowned information and problem-solving services. The resulting distributed, heterogeneous environment imposes certain requirements on the technical infrastructure. These requirements are identified, and a multi-agent problemsolving architecture is proposed to address them. 1. Introduction With the increasing complexity and globalization of society and commerce, many organizations are working to develop and deploy cooperating information systems (CIS). Such systems exhibit two key characteristics: (1) they work together, such that they are able to coordinate the...

Distributed heterogeneous databases consist of systems which differ physically and logically, containing different data models and data manipulation languages. Although these databases are independently created and administered they must cooperate and interoperate. Users need to access and manipulate data from several databases and applications may require data from a wide variety of independent databases. Therefore, a new system architecture is required to manipulate and manage distinct and multiple databases, in a transparent way, while preserving their autonomy. This report contains an extensive survey on heterogeneous databases, analysing and comparing the different aspects, concepts and approaches related to the topic. It introduces an architecture to support interoperability among heterogeneous database systems. The architecture avoids the use of a centralised structure to assist in the different phases of the interoperability process. It aims to support scalability, an...

Advanced object-oriented applications require the management of schema versions, in order to cope with changes in the structure of the stored data. Two types of versioning have been separately considered so far: branching and temporal. The former arose in application domains like CAD/CAM and software engineering, where different solutions have been proposed to support design schema versions (consolidated versions). The latter concerns temporal databases, where some works considered temporal schema versioning to fulfil advanced needs of other typical objectoriented applications like GIS and the multimedia ones. In this work, we propose a general model which integrates the two approaches by supporting both design and temporal schema versions. The model is provided with a complete set of schema change primitives for full-fledged version manipulation whose semantics is described in the paper. Keywords: Schema versioning, Schema evolution, OODBMS, Temporal databases 1 Introduction In th...

. In this paper, we discuss some aspects of database support for digital libraries. From a DL perspective, database systems, and in particular, object database systems provide a nice basis for future DL systems. More generally, database research provides solutions to many DL issues even if these are partial or fragmented. When possible, work should not be duplicated and good software and ideas should be reused. From a DB perspective, we want to stress that digital libraries propose beautiful applications and challenges to DBMS technology. They suggest a number of improvements to DBMSs that could be beneficial beyond DL applications. 1 Introduction The success of the Internet and the Web, and the popularity of Intranet information servers have brought Digital Libraries to the forefront of research. Digital libraries (DL) until recently have focused on issues such as data digitization, search in document collections, and relevance ranking. The scaling up of DL applications stresses new ...