On scientific workflow
Jinjun Chen, Wil M. P. van der Aalst · Swinburne figshare (Swinburne University of Technology) · 2007
Scientific workflow is another type of workflow that usually underlies in many complex e-science applications such as climate modeling and disaster recovery simulation. From the perspective of such e-science applications, based on Dieter Cybok’s paper in GGF10 there are at least three motivations for scientific workflow: 1) some complex e-science applications often require the creation of a collaborative workflow; 2) many e-scientists lack the necessary low-level expertise to utilize the current generation of underlying computing infrastructure such as Grid toolkits; 3) workflow specifications can be reused, modified and shared once they are defined. With these motivations, specific requirements for developing a scientific workflow management system need to be identified. The requirements may cover several aspects such as data or computation intensity and lifecycle management of participating services. The investigation of such requirements may be carried out from the following two points of view.