One sentence summary
- SCA should be (seen to be) leader in mainstream computing technologies for the lab, a reliable partner for collaboration and able to develop and support tools for the wider physics community.
First Steps
- Developing criteria/metrics for project support levels within SCA** Take into account need to support existing experiments/projects vs need to develop new initiatives** Support single experiment/projects vs set of SLAC based projects vs extended community
- Internal funding vs external funding
- Understanding existing and planned projects within SCA** Including projects which SCA could/should be involved in** Try to identify common requirements which can be "factored out"*** Maximize reuse, eliminate unnecessary/unintentional duplication
- Defining technical direction for new projects** Attempt to balance leveraging of existing skill-sets vs understanding and exploiting new industry trends** Need people's input on what we should be doing
- Form "task-force" to evaluate some new technologies
- Create SLAC wide developers "journal-club"
- Developing targets of opportunity for new projects and collaborations** Set of projects ready when new funding opportunities present themselves** Developing priority list for existing/new projects
- Creation of (new) project teams** Possibly bring people together from different parts of SCA, and people from outside of SCA** E.g. LSST Camera Control System ** SLAC Max, Tofigh, Owen Saxton, Stuart Marshall*** Paris, Indiana, Arizona, Brookhaven, ...
Possible Projects (very early ideas -- not meant to all be good ideas)
- New Projects** Projects with Computing Division** Support for GIT, Fisheye, ...** Interactive web based scientific computing status page
-
- Better integration between Crowd, Group Manager, Single sign-on
- Account management
- mathjax for confluence
- LSF with virtualization
- Data Management** Data Portal for Photon Science (and Astro physics)** Best features from Fermi data catalog, JCSG, CXIDB, ICAT, Dirac, ...** Why stop at presenting data location via the web?*** Should be able to drill into and operate on data*** In a collaborative way
- Web based scientific data visualization tools*** Extension/compatible with Google Visualization API
- Virtual Observatory integration
- LCLS real-time image displays
-
- Existing Projects
Impoved SCA web presence
Possible Collaborators (see above)
- Atlas group at Argonne (
Sergei Chekanov, David Malon, ...)
- Root team
- Tech-X
- Google/Amazon/...
Possible Technologies (see above)
- Technology selection criteria** Lifespan/stability** Adoption/Support
- Popularity/Mindset
- Language neutrality/Interoperability
- Google visualization API
- Web/HTML5/GWT/Dart/Json
- HDF5
- IPython/IPython Notebook
- Virtual Observatory
- MatLab
- DGAS (http://www.to.infn.it/dgas/)
- LSF virtualized cluster support
- Google Cloud
- Replacement for LSF (Steffen)