Catálogo de publicaciones - libros

In a modern enterprise, it is inevitable that different portions of the organization will use different systems to produce, store and search their critical data. Competition, evolving technology, mergers, acquisitions, and geographic distribution all contribute to this diversity. Only by combining these various systems can the enterprise realize the full value of the data they contain. Yet building new applications across these various information sources can be amazingly painful, forcing developers to discover what data is where, figure out what it means, and learn myriad different interfaces.

A major aspect of complex Enterprise Architectures is the integration of existing heterogeneous IT-systems in a business process oriented way. The paper starts with the definition of terms as Enterprise Architecture and process orientation. Based on an empirical study the paper shows that there is no significant business process orientation in information system integration projects. Among other reasons this is due to deficits in understanding and managing integration methods and technologies. Therefore the paper addresses the evaluation and comparison of relevant integration architectures as a first step to work on that issue. The paper differentiates individually coded interfaces, centralized hub&spoke and distributed approaches based on standardized interface descriptions (Service Oriented Architecture – SOA). The mentioned empirical study was extended by an action research based prototyping to assure a reliable evaluation and comparison of the three integration architectures. To make them comparable they have been implemented in the same fictitious business scenario which is described briefly. The paper finally compares the integration architectures with a set of 11 criteria which summarize over 400 variables taken into consideration in the evaluation process. The conclusion of the paper is not a solution but a suggestion for further research.

This paper proposes an approach to integrate a software product line (Aurora) with reflective rule-based business process modeling (RUMBA). Aurora is a service-oriented application development and execution platform supporting today’s well known “Rich Internet Applications” and “Enterprise Internet Applications” concepts. On the other hand, RUMBA is a rule-based model in which rules and rule-sets can be expressed in terms of dynamic aspects and delegated facts. The proposed approach mainly addresses “Reflective Aspect” and “Reflective Rule” patterns for the seamless integration of Aurora and RUMBA. Both architectural patterns introduce a “generative” approach for developing the basic aspects, dynamic rules and rule-sets so that all can be implemented in the Adaptive Object Model (AOM). The proposed model will be explained in detail and exemplified with existing projects using both Aurora and RUMBA approaches.

Enterprise-wide data replication improves availability, performance, fault-tolerance and dependability of database services within and between different subunits in medium and large enterprises. The overhead of consistency management of replicated data can be tamed by built-in DBMS functionality. Different kinds of applications, e.g., update-intensive online transaction processing, cyclical updates for data warehousing, knowledge sharing of repository data, and so on, have different requirements for the availability, up-to-dateness and consistency of replicated data. Thus, replication strategies should be adaptable to the specific requirements of diverse enterprise applications. We describe a middleware for enterprise-wide data replication. It maintains meta data for several protocols, so that the replication strategy can be adapted on the fly to the actual needs of an application.

Enterprise systems are mission critical. As a result, ensuring their correctness is of primary concern. This paper aids to the analysis of Web applications, focusing on the aspects related to the interaction of business logic and Web browsers. The method adopted is based on the Model Driven Architecture. First, the Platform Independent Model of Web applications is refined to create a new model called Abstract Description of Interaction (ADI). An ADI is a UML class diagram annotated with OCL statements to represent an abstraction of the interaction between the thin client and the business logic. Secondly, the ADI model is automatically transferred to an Alloy model and analysed using the Alloy Analyser.

This paper describes the EDIS business messaging architecture, which is a modern, lightweight system that is used in numerous companies. It explains the requirements for such a system, the problematic issues that have to be dealt with, and also some aspects of the wider context of e-commerce. Furthermore, it compares the presented architecture to similar systems like, for example, MS BizTalk and discusses related research on enterprise architecture.

In a distributed ERP (Enterprise Resource Planning) system, the different local ERP systems are integrated in such a way that each local system can use the resources/stocks managed by the other local ERP systems. Businesses with branch offices may derive great benefits from such systems. In theory, such a system can be built by using a distributed DBMS (Data Base Management System). However, distributed DBMSs are not used in practice as e.g. performance and local autonomy are low. In distributed databases with relaxed ACID properties (Atomicity, Consistency, Isolation, and Durability), it is possible to optimize performance, local autonomy, and availability by using short duration locking. In such systems, data is not locked across locations, i.e. data is not locked across a dialog with a user, and replicated data must be updated asynchronously. In this paper, we will describe how it is possible to design a distributed ERP system by using databases with relaxed ACID properties. The techniques described are general in the sense that most package software run on separate computers may be integrated by using the same methods. The author has cooperated with one of the major ERP software companies in analyzing how the company can design such a distributed version of their ERP system.

A data warehouse is filled with data during the extraction process. Such a process is sometimes interrupted by occurrence of a failure. After a failure the warehouse contains an incomplete data set, a part of the set is missing. To load the missing part of the data one of the interrupted extraction resumption algorithms is usually used. In this paper we analyze the influence of data balancing used in a distributed data warehouse on the efficiency of extraction and resumption processes. During resumption we base on the Design-Resume algorithm which imposes no additional overhead on an uninterrupted extraction process. We present how the balancing is done and examine its influence on the ETL process efficiency. Finally, basing on the results of performed tests, we discuss advantages and disadvantages of the balancing with respect to the ETL process.

OLAP applications are currently widely used in business applications. These applications are implicitly defined on top of OLTP systems. The applications make use of aggregation functions and data combinations. A number of paradoxes is observed if arbitrary aggregation functions and combinations are used. We develop a theory of aggregation functions, OLTP-OLAP transformations, and of the data cube. Based on these investigations we derive an that supports sound and correct querying: OLTP-OLAP specification frames. The specification frame of OLTP-OLAP schemata specifically emphasises soundness of all operations involved by built-in guards. Or to turn it around, we make provision that an innocent user does not start non-sense operations. This specification frame is based on OLTP schemata, OLTP-OLAP transformations, and a rigid theory of OLAP schemata and functions.

Todays data stream management systems (DSMSs) lack security functionality. Based on adversary scenarios we show how a DSMS architecture can be protected. We sketch a general DSMS architecture and introduce security issues that need to be considered. To face the threats we develop an extended system architecture that provides the necessary security mechanisms. We descuss the chosen concepts and illustrate how they can be realized by various system components. Our design focus is, considering the unique properties of data stream engines, to keep the impact on existing system components as little as possible and to limit the effect on the overall performance to a minimum.