Social Science Gateway to TeraGrid
@ Cornell University
What?
Usual social science data management and analysis tools not available on TeraGrid (TG) itself: SAS, SPSS, Stata
Also provides standard TG tools: R, compilers, etc.
Graphical desktop access to SSG compute nodes
High-speed (10Gbps) access to TeraGrid and thus to thousands of compute nodes and/or data
Local data storage (base allocation includes some storage, additional storage from Cornell storage cloud at cost)
TeraGrid
U.S. Census RDC
What?
More information
How?
Request TG
allocation
Request SSG account
Start using SSG
Ready for TG?
What else?
Other Cornell resources
Cornell TG Matlab cluster
Additional storage from Cornell storage cloud at cost
SSG = VirtualRDC
VirtualRDC replicates the environment of the U.S. Census RDC network (Linux OS, GUI)
Provides zero-obs datasets replicating data structure of RDC confidential data
Provides (where available) synthetic versions of confidential datasets, including early-release access

Who?
Researcher has existing theory/model/data
Project requires substantially more resources than typically locally available
Researcher is comfortable with the traditional data tools (SAS, SPSS)
Less comfortable with TeraGrid tools
Iterate
!
Until finished
TeraGrid? What is that?
You have this great idea...
But it requires substantial compute resources,
and your real expertise is with SAS/SPSS/Stata
(the usual social science tools)
Now you've received your 
TG allocation, and your TG 
credentials....
Now you have
TG access
Access to SSG and its tools
Now you need to transfer your data to the SSG
Log on to the graphical desktop,
"pull" or "push" your data to the SSG,
and start using SAS/SPSS/Stata
Your data has been prepped on the SSG,
your model has been tested, now you
push your model to the TeraGrid
for the "real" computation....
Oops... something not quite right...
back to the SSG...
After several iterations
between data prep stage 
and estimation stage (the usual...), you're done!

(RDC)
The Census RDC network provides secure and controlled access to  a large number of very detailed  (and thus confidential) datasets on the U.S. population and economy.