Workshop Scientific workflow and business workflow standards in eScience

1
WorkshopScientific workflow and business
workflow standards in e-Science

• Adam Belloum and Zhiming Zhao
• University of Amsterdam
• Virtual Laboratory for e-Science

2
Panel discussion

• Scientific workflows in VL-e
• Challenging issues in VL-e workflows
• Discussion

3
Workflows _at_ e-Science2006, Marian Bubak

• Wf a high-level paradigm for development of
  large, collaborative applications the only one?
• WfS PSE?
• WfS OS?
• WfS Grid / GRID ? (most WfS born before 1997)
• Computer science vs pure software engineering
• Low level paradigms (Petri nets, TM, SA, )
• Exploitation of knowledge (K-WfGrid and its
  sisters)
• Interactivity
• Interoperability of WfS (a new PL1) or a generic
  WfS with interfaces to
• Computing resources
• Data
• Domain ontologies
• (Provenance subsystem)
• Is it time for (standard) libraries of sci
  components / WS ? How?
• Lessons learnt from commercial world

4
Workflow support in VL-e

• How to move to e-Science?
• How to describe and run e-Science experiments?
• What model?
• What WF language?
• How to decompose functionality?

• How to choose a workflow system?
• One or more?
• How to deal with missing functionality?

Domain specific applications
These issues exist in almost all e-Science
projects, but they become more difficult when
different application domains are supported.
State of the art e-Science services.
Different levels of e-Science services.
Kepler, taverna, and triana
Existing workflow systems.
Workflow in VL-e effective and concrete solutions
for interfacing, Integration, and interoperability
How to choose a standard when do integration?
How to support human in the loop? -Provenance? -Fa
ult tolerance?
Grid infrastructure.
5
The vision of the panelists

• What will be the future of e-Science workflow
  management systems in a couple of years?
• What will be the top three issues to be
  addressed?

6
Themes in this workshop

• Running scientific experiments on e-Science
  environments
• Do we need a standard workflow language, instead
  of all the existing Languages (MoML, ScuFL )?
• Do e-Science applications need industrial
  workflow standards? Are industrial workflow
  standards suitable for e-Science applications?
• Will scientific applications need a standard?
  What should be standardized? Do they exist in
  industrial standards? Is BPEL a good candidate as
  standard workflow language?
• Sharing and integrating e-Science resources in
  experiments. Scientific applications were and
  still are the driving force for developing the
  Grid middleware. The last few year Grid
  middleware has shifted toward service oriented
  paradigm
• Effective exploration in scientific experiments.
  So far none of the proposed e-Science WMS has
  imposed itself as a de-facto standard for
  e-science.

7
Themes in this workshop

• Running scientific experiments on e-Science
  environments
• Sharing and integrating e-Science resources in
  experiments. Scientific applications were and
  still are the driving force for developing the
  Grid middleware. The last few year Grid
  middleware has shifted toward service oriented
  paradigm
• is this shift beneficial for the driving force?
  And what would be the impact of SOA on the
  development of the e-Science WMS?
• Do e-Science application scientists have to be
  service oriented?
• Will industrial service development tools fit in
  the scientific development environment and meet
• Effective exploration in scientific experiments.
  So far none of the proposed e-Science WMS has
  imposed itself as a de-facto standard for
  e-science.

8
Themes in this workshop

• Running scientific experiments on e-Science
  environments
• Sharing and integrating e-Science resources in
  experiments.
• Effective exploration in scientific experiments.
  So far none of the proposed e-Science WMS has
  imposed itself as a de-facto standard for
  e-science.
• Should we aim at one standard e-Science WMS? Or
  should we focus on making the existing WMS
  interoperate to achieve a typical e-science
  experiment?
• Will the diversity of scientific workflow systems
  be smaller after industrial standards? Will
  interoperability issues still exist? If so, what
  will they be?