Contextual Linking between Workflow Provenance and System Performance Logs

Ahanach, Elias el Khaldi; Koulouzis, Spiros; Zhao, Zhiming

doi:10.1109/escience.2019.00093

Cited by 6 publications

(5 citation statements)

References 9 publications

(9 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Metadata descriptions of research assets are not limited to 'characteristic' information; provenance data (which might be structured according to a standard such as PROV-O) for data products and processes are also an important target for semantic linking, especially for creating unified (or at least unifiable) records of how research assets are used and where they came from; such records may be generated from scientific workflow management systems with provenance support [35,42]. Such systems remain important for reproducible data science; most scientific investigations must follow a clear workflow, and there have been a number of workflow management systems developed with different characteristics and target applications [36], several of which have been applied to data science [37].…”

Section: Discussionmentioning

confidence: 99%

Semantic Linking of Research Infrastructure Metadata

Martin

Magagna

Liao

et al. 2020

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

The use of metadata to characterise scientific datasets, making data easier to discover and use directly by researchers and via various online data services, is one of the primary concerns of research infrastructures (RIs); also, of concern is the use of metadata to describe equipment, facilities, services and other research assets. Metadata models and terminology differ greatly between different communities and infrastructures however, and so make synthesising complex interdisciplinary scientific workflows involving assets from multiple RIs very challenging.'Semantic linking' addresses the need to enhance the interoperability of RI services and data by bridging metadata schemes, ontologies and vocabularies used by different research communities, whether by standardising the terminologies and schemes used by those communities, or by dynamically transforming metadata from one standard to another when retrieved by services on behalf of researchers executing their scientific workflows.Multiple techniques for and modes of semantic linking have been investigated in the context of the ENVRI community cluster of environmental and Earth science RIs, including top-down modelling of entities and activities within a standard reference model, enrichment of existing metadata records with shared terminology, full transformation of metadata records from one standard to another, and the generation of additional links to existing online data. We review some of these activities and their application to the promotion of semantic interoperability between RIs, and discuss other possibilities and recent developments that may also be useful for enhancing interdisciplinary data science.

show abstract

Section: Discussionmentioning

confidence: 99%

Semantic Linking of Research Infrastructure Metadata

Martin

Magagna

Liao

et al. 2020

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

show abstract

“…A workflow can be seen as a directed graph where each node is a function or module with input, output, and parameters, and each of the edges is a transfer of information [26] [27]. Elias et al [28] created a framework, cross-context workflow execution analyzer (CWEA), to connect provenance data with system logs and visualize the combination to aid scientists in detecting anomalies. Souza et al [29] used workflow provenance data to build a holistic view of the life cycle of scientific machine learning models.…”

Section: B Provenancementioning

confidence: 99%

Provenance-enhanced Root Cause Analysis for Jupyter Notebooks

Xin

Stallinga

Liu

et al. 2022

2022 IEEE/ACM 15th International Conference on Utility and Cloud Computing (UCC)

Self Cite

View full text Add to dashboard Cite

With Jupyter notebooks becoming more commonly used within scientific research, more Jupyter notebook-based use cases have evolved to be distributed. This trend makes it more challenging to analyze anomalies and debug notebooks. Provenance data is an ideal option that can create more context around anomalies and make it easier to find the root cause of the anomaly. However, provenance rarely gets investigated in the context of distributed Jupyter notebooks. In this paper, we propose a framework that integrates two data types, provenance and detected performance anomalies based on performance data. We use the combined information to visually show the enduser the provenance at the time of the anomaly and the root cause of the anomaly. We build and evaluate the framework with a notebook extended with anomaly-generating functions. The generated anomalies were automatically detected, and the combined information of provenance and anomaly creates a valuable subset of the provenance data around the time an anomaly occurred. Our experiments create a clear and confined context for the anomaly and enable the framework to find the root cause of performance anomalies in Jupyter notebooks.

show abstract

“…A Cross-context Workflow Execution Analyser (CWEA) is developed for users to effectively investigate possible workflow execution anomalies or bottlenecks by combining provenance with available system metrics [41]. The tool is able to retrieve available system logs of the particular machines (virtual machines if in Cloud) and align them with the provenance provided by the workflow management system.…”

Section: Provenance and System Logsmentioning

confidence: 99%

Data Provenance

Magagna

Goldfarb

Martin

et al. 2020

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

The provenance of research data is of critical importance to the reproducibility of and trust in scientific results. As research infrastructures provide more amalgamated datasets for researchers and more integrated facilities for processing and publishing data, the capture of provenance in a standard, machine-actionable form becomes especially important. Significant progress has already been made in providing standards and tools for provenance tracking, but the integration of these technologies into research infrastructure remains limited in many scientific domains. Further development and collaboration are required to provide frameworks for provenance capture that can be adopted by as widely as possible, facilitating interoperability as well as dataset reuse. In this chapter, we examine the current state of the art for provenance, and the current state of provenance capture in environmental and Earth science research infrastructures in Europe, as surveyed in the course of the ENVRIplus project. We describe a service developed for the upload, dissemination and application of provenance templates that can be used to generate standardised provenance traces from input data in accordance with current best practice and standards. The use of such a service by research infrastructure architects and researchers can expedite both the understanding and use of provenance technologies, and so drive the standard use of provenance capture technologies in future research infrastructure developments.

show abstract

Contextual Linking between Workflow Provenance and System Performance Logs

Cited by 6 publications

References 9 publications

Semantic Linking of Research Infrastructure Metadata

Semantic Linking of Research Infrastructure Metadata

Provenance-enhanced Root Cause Analysis for Jupyter Notebooks

Data Provenance

Contact Info

Product

Resources

About