Some research strands in workflow-based provenance

metadata about this mindmap

Originally prepared for: Dagstuhl seminar on Principles of Provenance, March, 2012

Author: P.Missier

Last updated: March 13, 2012

Modelling

The elusive database / process-based

provenance convergence:

"Is there a common notion of provenance

that can view wf provenance

and db provenance as different flavors

of the same broad model?"

IPAW'08

Kwasnikowska, Natalia, and Jan den Bussche. â€œMapping the NRC Dataflow Model to the Open Provenance Model.â€ In Provenance and Annotation of Data and Processes, edited by Juliana Freire, David Koop, and Luc Moreau, 5272:3-16. Springer Berlin / Heidelberg, 2008. http://dx.doi.org/10.1007/978-3-540-89965-5_3.

TAPP'10 - DFL

Acar, Umut, Peter Buneman, James Cheney, Jan Van Den Bussche, Natalia Kwasnikowska, and Stijn Vansummeren. â€"A graph model of data and workflow provenance"€ (February 22, 2010): 8. http://portal.acm.org/citation.cfm?id=1855795.1855803.

DFL (extension of NRC) for provenance

PVLDB'12 - the PigLatin paper

Olston, Christopher, Benjamin Reed, Utkarsh Srivastava, Ravi Kumar, and Andrew Tomkins. â€œPig latin: a not-so-foreign language for data processing.â€ In Proceedings of the 2008 ACM SIGMOD international conference on Management of data, 1099-1110. New York, NY, USA: ACM, 2008. http://doi.acm.org/10.1145/1376616.1376726.

Yael Amsterdamer 2 , Susan B. Davidson 1 , Daniel Deutch 3 , Tova Milo 2 , Julia Stoyanovich 1 , Val Tannen, Putting Lipstick on Pig: Enabling Database-style Workﬂow Provenance, PVLDB, vol. 5 n. 4, 2012

Breaks the module black box assumption and accounts for stateful modules

Key assumption: non-black box modules are encoded using Pig Latin

Ultimately grounded in NRC, as pig Latin expressions are translated into NRC

But this is not clear in the paper?

So that this can be used:
Y. Amsterdamer, D. Deutch, and V. Tannen. Provenance for aggregate queries. In PODS, 2011.

Contributions

A framework that marries database-style and workflow provenance models, capturing internal state as well as fine-grained dependencies in workflow provenance, when data operations are expressed using Pig Latin expressions.

A graph-based representation of fine-grained provenance for workflows, which also captures module invocations and module state changes.

Definition of graph transformation operations ZoomIn, ZoomOut and deletion propagation, which enable novel workflow analysis queries.

Queries over the provenance graphs limited to zooming in and out?

Implementation of the lipstick system

Joint modelling of execution traces

and process specification

the debugger argument: “I wouldn’t debug a running program without a link back to the source code”

the visualization argument: “the workflow structure is my roadmap into a trace with a complex unfolding”

baseline: OPM then PROV (W3C)

What is needed, really, and why?

Should these features have a place

in the provenance model?

data streaming data collections

(e.g. Taverna) and other data structures

concurrency of blocks execution

stateful computation

Is there a “killer app” that showcases the need

for workflow trace+structure provenance models?

Link between MoC and MoP

-- the “canonical MoP” for a MoC

T = R - I + M

Privacy-preserving provenance

position / early papers

WANDS'10 - Davidson et al.

Davidson, Susan B., Sanjeev Khanna, Sudeepa Roy, and Sarah Cohen Boulakia. “Privacy issues in scientific workflow provenance.” In First International Workshop on Workflow Approaches to New Data-centric Science (WANDS’10), edited by Paolo Missier, Vasa Curcin, and Susan Dadvidson. Indianapolis: ACM, 2010. http://portal.acm.org/citation.cfm?id=1833398.1833401.

CIDR'11 - Davidson et al.

Susan B. Davidson, Sanjeev Khanna, Sudeepa Roy, Julia Stoyanovich, Val Tannen, Enabling Privacy in Provenance-Aware Workflow Systems, Procs. CIDR'11

Position paper

trade-off: query utility vs privacy preservation

assumes queries span both workflow structure and execution trace

Data privacy -- standard in DBMS

Module privacy - function computed by a module should be (easily) inferrable

Structural privacy -- hides the fact that some modules contribute to the production of some data

SSDBM'11 - Dey, Zinn, et al.

Dey, Saumen, Daniel Zinn, and Bertram Ludäscher. “ProPub: Towards a Declarative Approach for Publishing Customized, Policy-Aware Provenance.” In Scientific and Statistical Database Management, edited by Judith Bayard Cushing, James French, and Shawn Bowers, 6809:225-243. Springer Berlin / Heidelberg, 2011. http://dx.doi.org/10.1007/978-3-642-22351-8_13.

PROPUB (Provenance Publisher)

state provenance publication and customization requests

specify provenance policies, check whether the policies are satisfied

repair policy violations and reconcile conflicts between user requests and provenance policies should they occur

ICDT'11 - Davidson et al.

Davidson, Susan B, Sanjeev Khanna, Sudeepa Roy, Julia Stoyanovich, Val Tannen, and Yi Chen. â€œOn provenance and privacy.â€ In Proceedings of the 14th International Conference on Database Theory, 3-10. New York, NY, USA: ACM, 2011. http://doi.acm.org/10.1145/1938551.1938554.

PODS'11 - Davidson et al.

Davidson, Susan B, Sanjeev Khanna, Tova Milo, Debmalya Panigrahi, and Sudeepa Roy. â€œProvenance views for module privacy.â€ In Proceedings of the thirtieth ACM SIGMOD-SIGACT-SIGART symposium on Principles of database systems, 175-186. New York, NY, USA: ACM, 2011. http://doi.acm.org/10.1145/1989284.1989305.

The secure-view problem.
Module privacy only

Explores the trade-off between the utility of provenance information and the privacy guarantees that authors/owners desire.

Collecting

Broading the scope: non-workflow RTEs

R is prominent in science, other scripting languages

Merging Provenance from different executions:

the virtual experiment

provenance is a manifestation of

Implicit collaboration through data sharing

across multiple workflows

the DataONE "Data Tree of Life" project

Missier, Paolo, Bertram Ludascher, Shawn Bowers, Manish Kumar Anand, Ilkay Altintas, Saumen Dey, Anandarup Sarkar, Biva Shrestha, and Carole Goble. "Linking Multiple Workflow Provenance Traces for Interoperable Collaborative Science."€ In Proc.s 5th Workshop on Workflows in Support of Large-Scale Science (WORKS), 2010.

Secure / tamper-evident /

trustworthy provenance for workflows

Police forces are becoming very interested in this

Storage security 2007

Hasan, Ragib, Radu Sion, and Marianne Winslett. â€œIntroducing secure provenance: problems and challenges.â€ In Proceedings of the 2007 ACM workshop on Storage security and survivability, 13-18. New York, NY, USA: ACM, 2007. http://doi.acm.org/10.1145/1314313.1314318.

position paper

Security/ usenix 2008

Braun, Uri, Avraham Shinnar, and Margo Seltzer. â€œSecuring provenance.â€ In Proceedings of the 3rd conference on Hot topics in security, 4:1--4:5. Berkeley, CA, USA: USENIX Association, 2008. http://dl.acm.org/citation.cfm?id=1496671.1496675.

The fake Picasso paper (TAPP 2009)

Hasan, Ragib, Radu Sion, and Marianne Winslett. â€œThe Case of the Fake Picasso: Preventing History Forgery with Secure Provenance.â€ In TAPP 2009, 1-14, 2009. http://www.usenix.org/events/fast09/tech/full_papers/hasan/hasan.pdf.

architecture for capturing prove-nance information at the application layer.

Scure and tamper-proof provenance chain. No-one can add or remove entries from the middle of the chain without detection

Interesting use cases where tampering with provenance chains may be desirable

The primary threat we guard against in this paper is undetected rewrites of history, which occur when ma-licious entities forge provenance chains to match illicit document writes and metadata modifications.

Multi layered approach: encryption for sensitive provenance chain record fields, a checksum-based approach for chain records and an incremental chained signature mechanism for securing the integrity of the chain as a whole.

Linked to issues of trustworthy computing infrastructure

trusted provenance @Tapp 10

Lyle, John, and Andrew Martin. Trusted computing and provenance: better together.€ In Proceedings of the 2nd conference on Theory and practice of provenance, 1. Berkeley, CA, USA: USENIX Association, 2010. http://dl.acm.org/citation.cfm?id=1855795.1855796.

Trusted Computing as a means to enforce secure provenance

Secure data management 2009

Zhang, Jing, Adriane Chapman, and Kristen LeFevre. â€œDo You Know Where Your Dataâ€™s Been? â€“ Tamper-Evident Database Provenance.â€ In Secure Data Management, edited by Willem Jonker and Milan Petkovic, 5776:17-32. Springer Berlin / Heidelberg, 2009. http://dx.doi.org/10.1007/978-3-642-04219-5_2.

The Tcloud EU project

Collecting and warehousing provenance:

PBase / GoldenTrail

"where can I find a repository of provenance traces that can mine?"

IDCC'11 conference on data curation

Missier, Paolo, Bertram Ludascher, Shawn Bowers, Ilkay Altintas, Saumen Dey, and Michael Agun. Golden Trail: Retrieving the Data History that Matters from a Comprehensive Provenance Repository.€ In Procs. 7th International Digital Curation Conference. Bristol,UK, 2011. http://www.dcc.ac.uk/events/idcc11.

Exploiting

Provenance for reproducibility

the wf4Ever EU project

emerging application: executable papers

Koop, David, Emanuele Santos, Phillip Mates, Huy T Vo, Philippe Bonnet, Bela Bauer, Brigitte Surer, et al. A Provenance-Based Infrastructure to Support the Life Cycle of Executable Papers. Procedia CS 4 (2011): 648-657.

Provenance for improving data engineering

Checkpointing facilities in support of long-running processes

partial/smart rerun of resource-intensive processes

progressive optimization of repetitive processes / self-tuning systems

TAPP'11

Missier, Paolo. â€œIncremental workflow improvement through analysis of its data provenance.â€ In Procs. TAPPâ€™11 (Theory and Practice of Provenance), Repeated executions of resource-intensive workflow. Heraklyion, Crete, Greece, 2011.