Catálogo de publicaciones - libros

In recent years, organisations have been undergoing a thorough transformation towards highly flexible and agile collaborations. Organisations are required to dynamically create and manage collaborative business processes to grasp market opportunities. Different from conventional business processes, a collaborative business process involves multiple parties and their business processes [1]. Thus, complex instance correspondences may exist at both build time and run time. Here, we choose to characterise instance correspondences in terms of cardinality and correlations. Thereby, we can define and represent statical and dynamic correspond-dence when modelling and executing a collaborative business process.

Process prioritization strategies, based on a probabilistic temporal model, are applied to reduce the number of deadline violations and the tardiness of workflows.

We present an abstract view of information integration based on a three dimensional (3D) space of Concepts, Data Models, and Domains. In this view, the first dimension specifies the concepts (e.g., entity sets, relations, classes, etc), the second dimension defines the data model in which a concept is represented (e.g., relational, semi-structured, object-oriented, etc), and the third dimension determines the concept domain which relative to the universe of the model, uniquely identifies the application domain. We also introduce three basic transformations, called X-transform, Y-transform, and Z-transform. The queries posed to the integrated level can be expressed on the basis of these basic queries.

Smart items technologies, like RFID and sensor networks, are the next big step in business process automation. We are currently developing a dedicated Smart Items Infrastructure (SII) that supports the development and deployment of distributed ubiquitous computing applications. Decomposing these applications into individual services and deploying them at different devices poses new technical challenges. In particular, we address service composition for heterogeneous device landscapes and mapping of service implementations to devices for execution.

RDF is a recommended standard for describing knowledge about resources over the Web. If we talk about querying RDF, we must consider important aspects concerning Web querying: distributed information and context oriented description of resources. In this paper we propose a framework that provides better, more complete answering to RDF queries than classical answering mechanisms. This framework provides a way to combine several RDF documents in order to compute a more complete answer to a given query. The combination of RDF documents is performed under some conditions, leading to a safe combination.

The increasingly complex and volatile nature of many contemporary markets means that demands are often best satisfied through dynamic networks of collaborating enterprises. Successful collaboration demands tight, flexible integration of business processes, however, this assumes that an appropriate team has been assembled. Traditionally, a toplevel service or goal is decomposed into component services or subgoals each of which is then matched to a provider. This is a complex task and while automated tools exist, supported especially by the notions of service discovery and traders, significant guidance is typically sought from the user. This imposes a substantial burden of interaction and considerable knowledge is demanded of a user to decompose to a level of detail which allows for matching to known services. Problems arise if this is not the case and open environments, such as the internet, present additional difficulties: if a user is not up-to-date potential decompositions may be missed; new entrants into a market may not be recognised; etc. Bottom-up approaches circumvent some of these difficulties, but also come at a price. For example, where goal decompositions are available, these are typically much more inefficient; if there is only a fixed number of processes available within a system, the case of “no solution” may take considerable time to establish. Moreover, since many bottom-up approaches distribute control, the system is vulnerable to malicious behaviour. Thus, a certain level of trust is required.

Composite Web Services (CWS) aggregate multiple Web Services (WS) in one logical unit to accomplish a complex task (e.g. business process). This aggregation is achieved by defining a workflow that orchestrates the underlying Web Services in a manner consistent with the desired functionality. Since CWS can aggregate atomic WS and/or other CWS they foster the development of service layers and reuse of already existing functionality. An important issue in the deployment of services is their run-time performance under various loads. Due to the complex interactions of the underlying services, a CWS they can exhibit problematic and often difficult to predict behaviours in overload situations.

This paper focuses on the use of request scheduling for improving CWS performance in overload situations. Different scheduling policies are investigated in regards to their effectiveness in helping with bulk arrivals.

Real time group editors allow two or more users at different locations to work on a shared document at the same time. In a mobile network environment with non-deterministic communication latency, a replicated architecture is usually adopted for the storage of the shared document in order to provide high responsiveness. A conflict occurs when two or more users have different intentions for editing the same part of the replicated document. Conflict can be categorised into two types: and conflicts. An conflict occurs when the conflicting operations cannot be realised at the same time, and if serially executed, the effect of the later operation will override the earlier operation. In contrast, a conflict occurs when the conflicting operations can be realised at the same time and both operations can be applied to the target without one overriding the other.

Distributed applications that adapt as their environment changes are developed from self-managing, self-configuring and self-optimising behaviours. This requires constant monitoring of the state of the environment, and analysing multiple sources of events. Event correlation is the process of correlating monitored events from multiple sources for further analysis. It is essential that event correlation supports reliable event management with minimal delay. This paper describes the design and implementation of an event correlation architecture for adaptive J2EE applications. The architecture supports flexible configuration of event correlation in terms of the reliability and performance. This is especially useful in situations when multiple sources of events have different level of requirements for reliability and performance. We evaluate the performance overhead of this event correlation architecture and demonstrate its practical usage in a case study of an adaptive image server application.

Peer-to-Peer computing has evolved over the last few years and is applied to a rising number of applications. Following this development we present a decentralised approach to dynamically select, load and integrate peer-to-peer based services into a CORBA-compliant middleware. This is achieved by extending and improving the mechanisms for dynamic service integration of JXTA an open peer-to-peer infrastructure. At object level we build on the fragmented object model provided by the AspectIX middleware to seamlessly integrate and use peer-to-peer services instead of common CORBA client/server-based implementations.