On 07/07/2011 10:25, Efthymia Tsamoura wrote:
> Hello

Hello

[snip of interesting service-invocation ordering]

> It would be very useful to know if there exist similar scientific
> workflow examples (where the order of task invocation can change and
> it is not known a-priori by the user, while the workflow performance
> depends on the workflow task invocation order)

There are certainly problems where different ordering of the services 
(or very similar services) can still achieve the same result but have 
very different performace. I know that for some bioinformatics workflows 
pre-filtering ids before you query a database can have a huge impact on 
performance.

I'm not sure what you mean by "the order of task invocation can change 
and it is not known a-priori by the user". For a specific workflow, 
services in Taverna are invoked as soon as the data is available (apart 
from some special cases). It depends on the workflow as to whether you 
can tell which order the services are invoked in. Basically, if the 
workflow branches then you cannot normally predict the relative order of 
service invocation in the different branches.

If you are asking if there are times when users cannot decide which 
order to put the services in the workflow, then there are almost 
certainly such cases.

> and if you are
> interested in extending taverna with optimization algorithms for such
> workflows.

It sounds interesting. Are you thinking of dynamically altering the 
order during an actual workflow run, or does this happen during workflow 
design? Even if you managed to just warn that a workflow was inefficient 
it would help a lot.

How do you model the semantics of the services?

> I am asking because i have recently developed an optimization
> algorithm for this problem and i would like to test its performance in
> a real-world workflow management system with real-world workflows.

I think it is best to copy your question to the Taverna users' mailing 
list as the subscribers may be able to point you at particular workflows.

Have you looked at http://www.myexperiment.org to see if there are any 
relatively simple workflows that you can test on?

> P.S.: references to publications or any other information dealing with
> scientific workflows of the above rationale will be extremely useful.

The one that I am most familiar with is the work done on the eLico 
project for generating optimized data mining workflows. It is at 
http://www.e-lico.eu/?q=DMOP

> Thank you very much for your time

Alan


------------------------------------------------------------------------------
All of the data generated in your IT infrastructure is seriously valuable.
Why? It contains a definitive record of application performance, security 
threats, fraudulent activity, and more. Splunk takes this data and makes 
sense of it. IT sense. And common sense.
http://p.sf.net/sfu/splunk-d2d-c2
_______________________________________________
taverna-hackers mailing list
[email protected]
Web site: http://www.taverna.org.uk
Mailing lists: http://www.taverna.org.uk/about/contact-us/
Developers Guide: http://www.taverna.org.uk/developers/

Reply via email to