Repository logo

Towards Collaborative Scientific Workflow Management System



Journal Title

Journal ISSN

Volume Title





Degree Level



The big data explosion phenomenon has impacted several domains, starting from research areas to divergent of business models in recent years. As this intensive amount of data opens up the possibilities of several interesting knowledge discoveries, over the past few years divergent of research domains have undergone the shift of trend towards analyzing those massive amount data. Scientific Workflow Management System (SWfMS) has gained much popularity in recent years in accelerating those data-intensive analyses, visualization, and discoveries of important information. Data-intensive tasks are often significantly time-consuming and complex in nature and hence SWfMSs are designed to efficiently support the specification, modification, execution, failure handling, and monitoring of the tasks in a scientific workflow. As far as the complexity, dimension, and volume of data are concerned, their effective analysis or management often become challenging for an individual and requires collaboration of multiple scientists instead. Hence, the notion of 'Collaborative SWfMS' was coined - which gained significant interest among researchers in recent years as none of the existing SWfMSs directly support real-time collaboration among scientists. In terms of collaborative SWfMSs, consistency management in the face of conflicting concurrent operations of the collaborators is a major challenge for its highly interconnected document structure among the computational modules - where any minor change in a part of the workflow can highly impact the other part of the collaborative workflow for the datalink relation among them. In addition to the consistency management, studies show several other challenges that need to be addressed towards a successful design of collaborative SWfMSs, such as sub-workflow composition and execution by different sub-groups, relationship between scientific workflows and collaboration models, sub-workflow monitoring, seamless integration and access control of the workflow components among collaborators and so on. In this thesis, we propose a locking scheme to facilitate consistency management in collaborative SWfMSs. The proposed method works by locking workflow components at a granular attribute level in addition to supporting locks on a targeted part of the collaborative workflow. We conducted several experiments to analyze the performance of the proposed method in comparison to related existing methods. Our studies show that the proposed method can reduce the average waiting time of a collaborator by up to 36% while increasing the average workflow update rate by up to 15% in comparison to existing descendent modular level locking techniques for collaborative SWfMSs. We also propose a role-based access control technique for the management of collaborative SWfMSs. We leverage the Collaborative Interactive Application Methodology (CIAM) for the investigation of role-based access control in the context of collaborative SWfMSs. We present our proposed method with a use-case of Plant Phenotyping and Genotyping research domain. Recent study shows that the collaborative SWfMSs often different sets of opportunities and challenges. From our investigations on existing research works towards collaborative SWfMSs and findings of our prior two studies, we propose an architecture of collaborative SWfMSs. We propose - SciWorCS - a Collaborative Scientific Workflow Management System as a proof of concept of the proposed architecture; which is the first of its kind to the best of our knowledge. We present several real-world use-cases of scientific workflows using SciWorCS. Finally, we conduct several user studies using SciWorCS comprising different real-world scientific workflows (i.e., from myExperiment) to understand the user behavior and styles of work in the context of collaborative SWfMSs. In addition to evaluating SciWorCS, the user studies reveal several interesting facts which can significantly contribute in the research domain, as none of the existing methods considered such empirical studies, and rather relied only on computer generated simulated studies for evaluation.



Collaborative, Scientific, Workflow, Management, System, Data, Analysis



Master of Science (M.Sc.)


Computer Science


Computer Science


Part Of