[seek-dev] SAS wrappers

Matt Jones jones at nceas.ucsb.edu
Mon Oct 13 10:16:12 PDT 2003


As Chad said, our previous system Monarch had support for SAS, Matlab, 
and R.  We should have actors for all of these in Ptolemy.  Chad and I 
have discussed a few issues, though.

One is that our previous code basically used the commandline to submit 
SAS jobs, so there was a large performance hit associated with process 
startup.  It would be better to have a service that already has one or 
more SAS sessions running to which jobs could be submitted.  So, we may 
consider developing a Grid Service that exposes SAS for licensed 
individuals and is accessible via a Ptolemy Grid Service actor.  This is 
not exclusive with developing a dedicated SAS actor that runs in batch 
mode, but would be much more efficient.  Similar issues will arise with 
R and Matlab and other services.

Another issue is that SAS as an actor is pretty different from other 
actors in Ptolemy.  In particular, as a general purpose analytical 
engine, modeling the SAS actor's inputs and outputs is done at a 
different level of abstraction.  It takes arbitrary data objects and sas 
code as inputs, and outputs any number of arbitrary data objects.  The 
SAS code is what defines the number and type of required data inputs and 
outputs.  So, we really need a way to indicate that a SAS actor (or a 
GridServie or Web Service actor) is dynamically reconfigurable in terms 
of the semantic types of the data it supports.  This is going to be an 
important issue in general in terms of the classification of analytical 
services on the EcoGrid.

Matt

Chad Berkley wrote:
> Yeah, the plan is to have an actor that can execute arbitrary sas code. 
>  We basically had this written for Monarch, so I just need to figure out 
> how to port it to a ptolemy actor.  We also want and actor for R. There 
> is already MatLab support in ptolemy, but I haven't looked into it 
> enough to see if the current actor will meet our needs.
> 
> chad
> 
> Bertram Ludaescher wrote:
> 
>> Chad:
>>
>> This is great! It answers my question. I guess this will ultimately
>> become a Ptolemy actor?
>> Bertram
>>
>>
>>>>>>> "CB" == Chad Berkley <berkley at nceas.ucsb.edu> writes:
>>>>>>
>>>>>>
>> CB> CB> Hey Bertram,
>> CB> We have not yet created a ptolemy SAS actor, if that's what you 
>> mean. CB> It's on my list to do though and I may try to get to it 
>> today.  We have CB> written execution plugins for SAS for our Monarch 
>> engine.  EML can CB> definitely document SAS datasets, if BIRN just 
>> wants to document them.
>> CB> If I didn't answer your question, let me know.
>> CB> CB> chad
>> CB> CB> Bertram Ludaescher wrote:
>>
>>>> Matt et al:
>>>>
>>>> Have you guys worked with SAS datasets and have a wrapper for SAS? 
>>>> For example as part of EML?
>>>> (context: some BIRN users have SAS data and we'd like to be able to
>>>> "wrap" sources)
>>>>
>>>> thanks
>>>>
>>>> Bertram
>>>> _______________________________________________
>>>> seek-dev mailing list
>>>> seek-dev at ecoinformatics.org
>>>> http://www.ecoinformatics.org/mailman/listinfo/seek-dev
>>>
>>>
>> CB> CB> CB> -- CB> -----------------------
>> CB> Chad Berkley
>> CB> National Center for
>> CB> Ecological Analysis
>> CB> and Synthesis (NCEAS)
>> CB> berkley at nceas.ucsb.edu
>> CB> -----------------------
> 
> 
> 

-- 
-------------------------------------------------------------------
Matt Jones                                     jones at nceas.ucsb.edu
http://www.nceas.ucsb.edu/    Fax: 425-920-2439    Ph: 907-789-0496
National Center for Ecological Analysis and Synthesis (NCEAS)
University of California Santa Barbara
Interested in ecological informatics? http://www.ecoinformatics.org
-------------------------------------------------------------------




More information about the Seek-dev mailing list