Beyond Scientific Workflows: Networked Open Processes

 

In the fast-evolving landscape of distributed scientific research, traditional approaches to building workflows are becoming unsustainable. With an ever-growing array of services and processes available, manually constructing workflows is no longer practical. Our research introduces a groundbreaking concept to revolutionize this domain: Networked Open Processes.

 

What Are Networked Open Processes?

 

Much like Linked Open Data transformed information sharing, Networked Open Processes bring the same philosophy to scientific computing. Processes are:

• Exposed: Made openly accessible for collaboration.

• Published: Documented for discovery and reuse.

• Linked: Connected through semantic relationships for seamless integration.

​​

This approach enables scientists to query and compose distributed processes dynamically, creating a more agile and scalable foundation for in silico experiments.

 

Key features

​

To realize this vision, we’ve developed:

​

1. Process Object Identifiers (POIs): Unique identifiers for referencing processes.

2. Semantic Function Templates: Reusable blueprints to standardize and simplify process definitions.

3. TReQL: A SQL-like language designed for querying interconnected process graphs, enabling intuitive exploration and integration of services.

 

PUMPKIN

 

At the core of our research is PUMPKIN, a cutting-edge framework that lays the groundwork for distributed data processing. PUMPKIN implements a protocol that facilitates scalable and efficient management of networked open processes, bridging the gap between diverse scientific services.
Explore more on the GitHub PUMPKIN repo.
 

 

[1] R. Cushing, M.T. Bubak, A.S.Z. Belloum, and C. de Laat, Beyond Scientific Workflows:Networked Open Processes, In Proceedings of the IEEE International Conference on e-Science 2013, workshop on Analyzing and Improving Collaborative eScience with SocialNetworks, doi:10.1109/eScience.2013.51.

​