Why-Diff: Explaining differences amongst similar workflow runs by exploiting scientific metadata

Thavasimani, Priyaa; Cała, Jacek; Missier, Paolo

doi:10.1109/bigdata.2017.8258275

Cited by 2 publications

(2 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It is supposed to work given a well-defined set of possible workflow edits and under assumptions regarding the completeness of the provenance graphs, and the availability of data diff functions. In earlier work [7], [15], we studied the simple case where only C and D may change, but not the structure of the programs W , W , and thus the provenance traces generated by E, E , when they are represented as provenance graphs, are isomorphic. Here we extend our previous study to cover the general case where E is obtained from E through a sequence of edits, and as a consequence, the traces collected at runtime are non-isomorphic and reflect structural differences introduced through these edits.…”

Section: B Contributionsmentioning

confidence: 99%

See 1 more Smart Citation

Why-Diff: Exploiting Provenance to Understand Outcome Differences From Non-Identical Reproduced Workflows

2019

Self Cite

View full text Add to dashboard Cite

Data analytics processes such as scientific workflows tend to be executed repeatedly, with varying dependencies and input datasets. The case has been made in the past for tracking the provenance of the final information products through the workflow steps, to enable their reproducibility. In this paper, we explore the hypothesis that provenance traces recorded during execution are also instrumental to answering questions about the observed differences between sets of results obtained from similar but not identical workflow configurations. Such differences in configurations may be introduced deliberately, i.e., to explore process variations or accidentally, typically as the result of porting efforts or of changes in the computing environment. Using a commonly used workflow programming model as a reference, we consider both structural variations in the workflows as well as variations within their individual components. Our why-diff algorithm compares the graph representations of two provenance traces derived from two workflow variations. It produces a delta graph that can be used to produce human-readable explanations of the impact of workflow differences on observed output differences. We report with Neo4j graph database. Further, we report explanations of the difference between workflow results using a suite of synthetic workflows as well as real-world workflows. INDEX TERMS Why-diff, provenance, reproducibility, big data, eScience Central, workflow.

show abstract

Section: B Contributionsmentioning

confidence: 99%

“…III-D and the delta graph presented in Sec. III-E. Why-diff extends the algorithm described in [15], which was limited to the simple case of a change in activity properties. At the top level, the algorithm consists of three phases (see Alg.…”

Section: F Proposed Algorithm: Why-diffmentioning

confidence: 99%