Leveraging business workflows in distributed systems research for the orchestration of reproducible and scalable experiments

Abstract

While rapid research on distributed systems is observed, experiments in this field are often difficult to design, describe, conduct and reproduce. By overcoming these difficulties the research could be further stimulated and add more credibility to results in distributed systems research. The key factors responsible for this situation are technical (software bugs and hardware errors), methodological (incorrect practices), as well as social (reluctance to share work). In this paper, the existing approaches for the management of experiments on distributed systems are described and a novel approach using business process management (BPM) is presented to address their shortcomings. Then, the questions arising when such approach is taken, are addressed. We show that it can be a better alternative to the traditional way of performing experiments as it encourages better scientific practices and results in more valuable research and publications. Finally, a plan of our future work is outlined and other applications of this work are discussed.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 7 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.