A layered workflow knowledge Grid/P2P architecture and its models for future generation workflow systems

摘要

This paper addresses the scalability issue on workflow systems using the knowledge Grid/P2P computing environment. As a feasible solution for the scalability issue, we design a Layered Workflow Knowledge Grid/P2P Architecture. The architecture is based upon a conceptual and generic framework that considers various perspectives for creating workflow knowledge Grid/P2P architectures. In order to show how the architecture is applicable in making up a future generation workflow system, we propose three types of workflow knowledge models and a physical architecture that is possibly configurable from the layered workflow knowledge Grid/P2P architecture. That is, the architecture has a 3-layered hierarchical configuration - procedure-layer, role-layer, and actor-layer - and the three workflow knowledge models - ICN-based model, role-based model, and actor-based model - are the workflow knowledge representations on Grid/P2P nodes in the corresponding layers, respectively. In particular, the most critical rationale of this paper is based on the fact that the nature of the knowledge Grid/P2P computing environment ought to be fitted very well into building a platform for future generation workflow systems characterized by maximally parallel and very large scale workflow procedures that are frequently observed in very large scale enterprizes. And we strongly believe that the rationale must be acceptable. Conclusively, the architecture and the models proposed in this paper shall be useful for not only maximizing the usability of computing facilities in the enterprize but also appropriately handling the scalability issue of its underlined future generation workflow management system in order to cope with massively parallel and very large scale workflow applications.