Javier Fabra

Controversy Corner: Automatic execution of business process models: Exploiting the benefits of Model-driven Engineering approaches

The business goals of an enterprise process are traced to business process models with the aim of being carried out during the execution stage. The automatic translation from these models to fully executable code which can be simulated and round-trip engineered is still an open challenge in the Business Process Management field. Model-driven Engineering has proposed a set of methodologies with which to solve the existing gap between business analysts and software developers, but the expected results have not as yet been achieved. In this paper, a new approach to solve this challenge is proposed. This approach is based on the integration of SOD-M, a model-driven method for the development of service-oriented systems, and DENEB, a platform for the development and execution of flexible business processes. SOD-M provides business analysts with a methodology that can be used to transform their business goals into composition service models, a type of model that represents business processes. The use of the Eclipse Modelling Framework and the ATLAS Transformation Language allows this model to be automatically transformed into a DENEB workflow model, resulting in a business process that is coded by a class of high-level Petri-nets and is directly executable in DENEB. The application of the proposal presented herein is illustrated by means of a real system related to the management of medical images.

RLinda: a petri net based implementation of the linda coordination paradigm for web services interactions

The core functionality of Web-service middlewares tries to wrap existing business logics and make them accessible as Web services. Recently, well-known standardization initiatives have proposed some high-level declarative languages for the description of coordination protocols and the implementation of coordination middlewares. In parallel to these initiatives, an increasing interest on the use of classical coordination models on distributed environments has been shown. In this work we present a Linda-like coordination framework using Petri nets, which is executed by the Renew tool, a high-level Petri net interpreter developed in Java, and subsequently exposed as a Web service able to be used by other services for coordination purposes. The implementation is based on an extension of the original Linda model that improves the tuple representation capabilities and extends the matching functions used for the recovery of tuples from the coordination space. The efficiency of the proposed implementation has been empirically evaluated on a cluster computing environment, and its performances compared with the previously reported ones related to JavaSpaces.

A framework for the development and execution of horizontal protocols in open BPM systems

A new generation of open Business Process Management (BPM) systems based on the service-oriented architecture and Web service technologies has recently emerged. The general tendency for these systems should be governed by the integration of independent Web-service specifications. Web services requirements guide the description, execution and choreography of business process and the implementation of frameworks for supporting the coordination, synchronization and creation of business transactions. However, a wide variety of open research issues related to the lack of maturity of the involved specifications makes the development of standard-based BPM systems difficult. In this paper we propose an abstract architecture inspired by Web service specifications to overcome these difficulties. Also, a particular implementation based on the Nets-within-Nets paradigm and the Renew tool is presented. The result is an executable infrastructure able to run business processes (their workflows and coordination protocols) as well as the horizontal protocols that guarantee a coherent outcome of their whole execution, such as the WS-Atomic Transaction protocol.

DRLinda: a distributed message broker for collaborative interactions among business processes

Recently, coordinationmiddleware systemshave evolved in order to describe coordination protocols in business process scenarios. This evolution proposes the use of three main components, being one of them a message broker to handle collaborative interactions among business processes. In a previous work, we proposed a framework for coordination in open BPM systems which used a centralised Linda-based implementation of amessage broker. The use of a centralised implementation leads to some common problems which a distributed model tries to solve in an efficient manner. In this paper, we present DRLinda, a distributed and dynamic implementation of the message broker based on the RLinda model, which improves and extends the RLindas features and can be configured at run-time, being suitable formore complex and highly-dynamic business process scenarios. The performance of the proposed implementation is empirically evaluated on a cluster computing environment.

Model Checking Analysis of Semantically Annotated Business Processes

Semantic business processes require new analysis techniques able to deal with behavioral properties that also consider semantic aspects. In this paper, a model checking method is introduced including semantic aspects in both the model description and the formula to be verified. In addition, Unary resource description framework (RDF) annotated Petri net systems, a formalism that allows the semantic description of business processes using RDF annotations, is formally defined and used to represent the input model of the model checker. Finally, the prototype implementations of both the Unary RDF annotated Petri net formalism and a model checker framework based on the use of RDF and SPARQL tools are also presented.

BPEL2DENEB: Translation of BPEL Processes to Executable High-Level Petri Nets

The use of the Service-Oriented Architecture paradigm (SOA) as an architectural model together with the use of Web services technologies have spread out widely in the world of business processes. The Business Execution Language, BPEL, represents one of the most common orchestration languages used to build SOA compositions. Using BPEL, a business process can be constructed integrating different Web services in the same process flow. However, BPEL is rather static and several approaches have been proposed in order to allow translating it to a set of different formalisms, being Petri nets one of the most extended. The main lack of these translations is that the result is also static and normally it is used merely to analyze, verify and study some behavioral and structural properties of the corresponding processes. In this work, we present the translation of BPEL processes to executable High-level Petri nets. The resulting nets are much more expressive and keep the original semantics and behavior than the original BPEL process. Moreover, these nets can not only be analyzed, but also executed directly by using the result as input in the DENEB platform for the development and execution of dynamic Web processes.

Runtime Protocol Binding: Flexible Service Integration by Means of Flexible Service Interactions

Dynamic and flexible service composition and interactionare a must in service oriented computing (SOC) scenarios. In this paper the authors present an extension to a previous work, the DENEB platform for the development and execution of Web processes, allowing Web processes to acquire and execute new interaction protocols at runtime. This makes DENEB a well adapted and flexible platform for service integration, useful to work in complex and evolving inter-organisational environments. The proposed solution is exemplified by means of the development of a real use case.

Cost Evaluation of Migrating a Computation Intensive Problem from Clusters to Cloud

Cloud has emerged as an alternative to clusters and grids. Its adoption as an execution environment capable of supporting the high requirements of scientific computations is still an open question. In a previous work, the authors conducted successfully a practical experience of taking advantage of clusters and grids to solve a semantic annotation problem in 178 days. In this work, the authors analyse the cost of solving that problem and compare it with the cost of solving it in a pure Cloud scenario. For this last, a detailed cost estimation is conducted according to the data extracted from the actual execution of a reduced dataset on Amazon EC2. As a result, the suitability of using Cloud-based solutions to solve large and complex scientific problems is discussed.

Enabling the Evolution of Service-Oriented Solutions Using an UML2 Profile and a Reference Petri Nets Execution Platform

The activities developed by a company (business processes) have to change frequently to adapt to the environment. The implementation of business processes should support these changes without any receding. In this work, we provide with an approach for modelling and executing agile and adaptable business processes. Our approach is based on UML2 separating choreography (stable interaction patterns) and orchestration (implementation of the evolving business process, also called workflows), allowing the transformation and execution of the models by means of a flexible SOA-based dynamic platform based on reference Petri nets.

Solving the Interoperability Problem by Means of a Bus

In the last years, many institutions have provided themselves with cluster and Grid infrastructures either for intensive computation or research objectives. Each infrastructure having its own and different management operating software, the integration of different platforms becomes a hard and complicated task. Solving the interoperability problem for a set of different computing infrastructures belonging to our institution in order to solve a computation intensive problem was our objective. The paper describes the solution that was applied and the experimental results obtained. The solution uses a platform based on a central bus shared by the involved system components for information exchange. In order to ensure that all computations will succeed, the solution includes cloud infrastructures to deal with situations in which the local computing resources pose some problems. Also a cloud based solution for the bus deployment is explored and empirically compared with a local deployment.