Distributed workflow supports collaborative processes composed of activities, in which the synchronization and coordination of the activities and the people having a part in them (the stakeholders of the process) is an essential characteristic. Workflow is based on the concept of a process model,describing the process to be followed, and on facilities (collectively termed the process enactment engine) for supporting and guiding the work of stakeholders according to that model. Distributed workflow technology nowadays provides many of the necessary paradigms, techniques and tools to support the management of complex, dynamic and decentralized business practices, their stakeholders, and their processes. There are a number of dimensions concurring to workflow distribution, which interact - and sometimes conflict - with one other in various, complex ways, such as distribution of the workflow information, the workflow actors, tools employed to carry out the work, and the work itself Complete decentralization of WF along all of its distribution dimensions is still a challenge. On the other hand approaches and techniques that have been established for and have become typical of distributed workflow show great potential towards the coordination of various kinds of distributed software systems, which require the execution of some kind of process: two important domains are the run-time monitoring and control of functional and non-functional properties of the components of a distributed system, and the dynamically determined cooperation within a group of software agents towards some common goal. This proposal intends to investigate the characteristics of those software coordination problems and how they can be tackled with distributed workflow models and techniques. We intend to carry out our investigation in the light of a case study directed towards process-aware provision of multimedia services and information to groups of dispersed users carrying out teamwork, including the automated workflow-based management of an advanced architecture for broadcasting and streaming video and audio, on the basis of available network and system resources. The case study can be also seen as a particular example within a more generic distributed system scenario,i.e. the continual validation of generic distributed applications. The scale of distribution is that of global dispersion of processes, software components,stakeholders, and data. Thus, we are going to experiment with a set of techniques and mechanisms, which address those aspects at the global networking scale, and whose integration within a globally decentralized workflow management system will help reconciling the various distributed dimensions of workflow, as well as resolving diverse distributed coordination problems.
展开▼
