Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revision Previous revision
Next revision
Previous revision
workshop-20100323 [2010/03/18 10:17]
johan
workshop-20100323 [2010/03/23 15:49] (current)
johan
Line 44: Line 44:
   * [[http://​www.isi.edu/​~deelman|Ewa Deelman]] and [[http://​graal.ens-lyon.fr/​~rbolze|Raphaël Bolze]], USC   * [[http://​www.isi.edu/​~deelman|Ewa Deelman]] and [[http://​graal.ens-lyon.fr/​~rbolze|Raphaël Bolze]], USC
   * Our partners of the [[http://​gwendia.polytech.unice.fr|GWENDIA projet]] (Grid Workflow ENactment for Data Intensive Applications). In particular:   * Our partners of the [[http://​gwendia.polytech.unice.fr|GWENDIA projet]] (Grid Workflow ENactment for Data Intensive Applications). In particular:
-    * Modalis: [[http://​www.i3s.unice.fr/​~johan|J. Montagnat]],​ [[http://​www.i3s.unice.fr/​~lingrand|D. Lingrand]], [[http://​users.polytech.unice.fr/​~hrenard|H. Renard]], [[http://​modalis.polytech.unice.fr/​~ketan|K. Maheshwari]] ​i(nvolved ​in workflow language definition and operation)+    * Modalis: [[http://​www.i3s.unice.fr/​~johan|J. Montagnat]],​ [[http://​www.i3s.unice.fr/​~lingrand|D. Lingrand]], [[http://​users.polytech.unice.fr/​~hrenard|H. Renard]], [[http://​modalis.polytech.unice.fr/​~ketan|K. Maheshwari]] (involved ​in workflow language definition and operation)
     * LIP laboratory: [[http://​graal.ens-lyon.fr/​~desprez|F. Desprez]], [[http://​graal.ens-lyon.fr/​~ecaron|E. Caron]] (involved in DAGs scheduling on a grid infrastructure)     * LIP laboratory: [[http://​graal.ens-lyon.fr/​~desprez|F. Desprez]], [[http://​graal.ens-lyon.fr/​~ecaron|E. Caron]] (involved in DAGs scheduling on a grid infrastructure)
     * CREATIS laboratory [[http://​www.creatis.insa-lyon.fr/​site/​en/​users/​glatard|T. Glatard]] (involved in medical application analysis pipelines porting and optimization on grids)     * CREATIS laboratory [[http://​www.creatis.insa-lyon.fr/​site/​en/​users/​glatard|T. Glatard]] (involved in medical application analysis pipelines porting and optimization on grids)
Line 51: Line 51:
 === How? === === How? ===
  
-The idea of the workshop is to identify what complementarities and overlap our teams have, what are each group strengths and weaknesses, and how this could be used in the future to have a fruitful collaboration. To allocate time to discussions,​ talks are scheduled with short presentations (15') and long discussions (25') on purpose.+The idea of the workshop is to identify what complementarities and overlap our teams have, what are each group strengths and weaknesses, and how this could be used in the future to have a fruitful collaboration. To allocate time to discussions,​ talks are scheduled with short presentations (25') and long discussions (35') on purpose.
  
 === What? === === What? ===
Line 62: Line 62:
   <li class="​level1"><​div class="​li">​ <​b>​Abstract workflow languages</​b>​ (languages, expressiveness,​ control and data flows)<​br>​   <li class="​level1"><​div class="​li">​ <​b>​Abstract workflow languages</​b>​ (languages, expressiveness,​ control and data flows)<​br>​
 Johan (MODALIS)<​br>​ Johan (MODALIS)<​br>​
-<​i>​GWENDIA,​ A data-driven workflow language for grids based on array programming principles</​i><​br>​+<​i>​GWENDIA,​ A data-driven workflow language for grids based on array programming principles</​i> ​(<a href="/​_media/​i3s_workflows-workshop-100323.pdf?​id=workshop-20100323&​amp;​cache=cache">​slides</​a>​)<br>
 <a href="​http://​rainbow.polytech.unice.fr/​publis/​montagnat-isnard-etal:​2009.pdf">​WORKS'​09 paper</​a></​div></​li>​ <a href="​http://​rainbow.polytech.unice.fr/​publis/​montagnat-isnard-etal:​2009.pdf">​WORKS'​09 paper</​a></​div></​li>​
  
   <li class="​level1"><​div class="​li">​ <​b>​Concrete workflow generation</​b>​ (executable workflows production and languages, meta-workflows,​ just-in-time production)<​br>​   <li class="​level1"><​div class="​li">​ <​b>​Concrete workflow generation</​b>​ (executable workflows production and languages, meta-workflows,​ just-in-time production)<​br>​
 <ul> <ul>
-<li class="​level1"><​div class="​li">​Ewa (USC)Pegasus<​br>​+<li class="​level1"><​div class="​li">​Ewa (USC)<​br>​ 
 +<i>Pegasus ​Workflow Management System</​i>​ (<a href="/​_media/​deelman_nice-3-2010.pdf?​id=workshop-20100323&​amp;​cache=cache">​slides</​a>​)<br>
 <a href="​http://​pegasus.isi.edu/​publications/​2009/​CallaghanS-ScalingUp.pdf">​Scaling up Workflow-based Applications (JoGC)</​a>, ​ <a href="​http://​pegasus.isi.edu/​publications/​2009/​CallaghanS-ScalingUp.pdf">​Scaling up Workflow-based Applications (JoGC)</​a>, ​
 <a href="​http://​pegasus.isi.edu/​publications/​2008/​bharathi-works08.pdf">​Characterization of Scientific Workflows (WORKS'​08)</​a></​div></​li>​ <a href="​http://​pegasus.isi.edu/​publications/​2008/​bharathi-works08.pdf">​Characterization of Scientific Workflows (WORKS'​08)</​a></​div></​li>​
Line 75: Line 76:
   <li class="​level1"><​div class="​li">​ <​b>​Workflow execution environment</​b><​br>​   <li class="​level1"><​div class="​li">​ <​b>​Workflow execution environment</​b><​br>​
 Tristan (CREATIS)<​br>​ Tristan (CREATIS)<​br>​
-<i>A review of operational nightmares and related dreams of grid workflows</​i>​ (<a href="/​_media/​nightmares.pdf">​preliminary ​slides</​a>​)+<i>A review of operational nightmares and related dreams of grid workflows</​i>​ (<a href="/​_media/​nightmares.pdf">​slides</​a>​)
 <ul> <ul>
   <li class="​level1"><​div class="​li">​ <a href="/​_media/​hg10.pdf">​Application facts</​a>​ (submitted to HealthGrid 2010)</​div>​   <li class="​level1"><​div class="​li">​ <a href="/​_media/​hg10.pdf">​Application facts</​a>​ (submitted to HealthGrid 2010)</​div>​
Line 91: Line 92:
  
   <li class="​level1"><​div class="​li">​ <​b>​Resource allocation and tasks scheduling</​b>​ (tasks management)<​br>​   <li class="​level1"><​div class="​li">​ <​b>​Resource allocation and tasks scheduling</​b>​ (tasks management)<​br>​
-Miron (UW), CONDOR</​div></​li>​+Miron (UW)<​br>​ 
 +<​i>​Can DAGs do it all - capturing data dependencies with DAGs</​i>​</​div></​li>​
 </ul> </ul>
 </​html>​ </​html>​
  
-These are 6 talks (15 minutes presentation + 25 minutes discussion = 40 minutes per talk -> less than 5h). Each speaker should address a list of questions (to be determined) to help identifying challenges, limitations and what achievement is targeted. We have about 3h left for general discussions (complementarities,​ expectations,​ cross-challenges,​ etc). On the second day we could elaborate on possible interaction and collaboration organization.+These are 6 talks (25 minutes presentation + 35 minutes discussion = 6h total). Each speaker should address a list of questions (to be determined) to help identifying challenges, limitations and what achievement is targeted. We have about 3h left for general discussions (complementarities,​ expectations,​ cross-challenges,​ etc). On the second day we could elaborate on possible interaction and collaboration organization.
  
 === French-US collaboration opportunities === === French-US collaboration opportunities ===