Robust Scientific Workflows with Service Automata

  
Scientific disciplines such as biology, astronomy, chemistry are becoming
extremely computationally intensive. An example of such a computationally
intensive system is the Pan-STARRS1 telescope [1]. The 1.4 gigapixel camera
produces a 3 GB image for every 30 seconds of exposure [2]. To manage the
complexity, workflows [2] are essential to produce objects usable by scientists.

Service Automata [3], is a novel framework for runtime enforcement of security
in distributed systems. The Service Automata framework was implemented in Java
for Java programs. This implementation has been extended to BPEL, that is a popular
framework for specifying and implementing workflows. In this talk, the workflows
used for the PS1 system will be simulated with BPEL and the models explained.
Then, we will explore the data integrity  requirements of published data products
and security requirements of the system with the help of the BPEL models.

The talk will contain a demo of the tool, focused on enforcing data integrity of the
published data products of the PS1 workflows modelled in BPEL, highlighting the
automatic, dependable instrumentation of workflows.

[1] http://ps1sc.org/
[2] Yogesh Simmhan, Catharine van Ingen, Alexander S. Szalay, Roger S. Barga, Jim Heasley:
        Building Reliable Data Pipelines for Managing Community Data Using Scientific Workflows. eScience 2009: 321-328
[3] Richard Gay, Heiko Mantel, Barbara Sprick: Service Automata. Formal Aspects in Security and Trust 2011: 148-163