Database Management Systems (DBMSs) that can be tailored to specific requirements offer the potential to improve reliability and maintainability and simultaneously the ability to reduce the footprint of the code base. If the requirements of an application change during runtime the DBMS should be adapted without a shutdown. Runtime-adaptation is a new and promising research direction to dynamically change the behavior of a DBMS. Especially the adaptation of the Transaction Manager (TM) states a challenge. In this paper, we present the session semantics of a runtime-adaptable TM. We define preliminaries and assumptions to activate the TM during sessions from a conceptual point of view. The advantages and disadvanteges of different ap-proaches are discussed, especially regarding the occurence of ANSI SQL phenomena. From a technical point of view, we define requirements for the architecture of the TM and the DBMS that arose in our prototype.

Databases play a key role in an increasingly diverse range of applications and settings. New requirements are continually emerging and may differ substantially from one domain to another, sometimes even to the point of conflict. To address these challenges, database systems are evolving to cater for new application domains. Yet little attention has been given to the process of researching and developing database concepts in response to new requirements. We present a platform designed to support database research in terms of experimentation with different aspects of database systems ranging from the data model to the distribution architecture. Our platform is based on the notion of metamodel extension modules, inspired by proposals for adaptive and configurable database management systems. However, rather than building a tailored system from existing components, we focus on the process of designing new components. To qualitatively evaluate our platform, we present a series of case studies where our approach was used successfully to experiment with concepts designed to support a variety of novel application domains.

Abstract. A focus of Grid computing are data intensive applications. Additionally, database management systems (DBMS) are gaining on importance in many scientific disciplines for publication of research results. The employment of Service-oriented-Architecture (SoA) raises the question of how DBMSs and their built-in technologies can be best utilized in such environments. A common way is to pull out all required data for a certain task from a source and process it service side far away from the original source. This approach is characterized by a passive usage of the DBMS as a pure data provider which implies significant overheads. The research effort described in this paper allows an active usage of a DBMS by relocating distributed query processing functionality inside it. Our novel solution utilizes the existing database technology, puts just the interface code at the service level while the data processing code resides at the database level and uses a push mechanism for the result data. The advantages are less overheads and data movement as well as increased data locality. Our proof of concept implementation is evaluated by comparing our distributed query processing prototype working inside popular relational DBMS (Oracle 10g and PostGreSQL 8.3) with a traditional installation of the OGSA-DQP middleware developed for distributed query processing on the Grid. 1

Abstract not found