As distributed computing infrastructures evolve, and as their take up by user communities is growing, the importance of making different types of infrastructures based on a heterogeneous set of middleware interoperable is becoming crucial. This PhD submission, based on twenty scientific publications, presents a unique solution to the challenge of the seamless interoperation of distributed computing infrastructures at the level of workflows.
The submission investigates workflow level interoperation inside a particular workflow system (intra-workflow interoperation), and also between different workflow solutions (inter-workflow interoperation). In both cases the interoperation of workflow component execution and the feeding of data into these components workflow components are considered.
The invented and developed framework enables the execution of legacy applications and grid jobs and services on multiple grid systems, the feeding of data from heterogeneous file and data storage solutions to these workflow components, and the embedding of non-native workflows to a hosting meta-workflow. Moreover, the solution provides a high level user interface that enables e-scientist end-users to conveniently access the interoperable grid solutions without requiring them to study or understand the technical details of the underlying infrastructure. The candidate has also developed an application porting methodology that enables the systematic porting of applications to interoperable and interconnected grid infrastructures, and facilitates the exploitation of the above technical framework.