SVAC Prototype Database
Ashmi Ashokkumar Chokshi' , Jian Zhong, Xin Chen ,
Dylan Thurston , Eduardo do Couto e Silva
Last update October 22, 2004
Check our new prototype and the old in case you like history
Overview
The primary calibration database is maintained by the Science Analysis Software (SAS) group since the reconstruction programs need to read in the calibration constants for the various detectors. In the current prototype, the SVAC database will extract, via scripts, data from the SAS database and populated tables of the SVAC database, which are designed to allow trend analysis with a pre-defined query system using a web interface.
SVAC
- Requirements
Allowing users to access data from the results GLAST calibration tasks and also allow them to perform
trend analysis on these data. To be made accessible via Internet/intranet.
How
the system will work
· As shown in Fig.1, our web pages will reside on the NT server. Users will access these web pages to view the data in various forms (histograms, tables etc.)
· Our data is on the Unix server, and we use Java Servlets to access the raw data and process it.
· Thus, the web page will invoke a servlet on the Unix machine, and this servlet will run inside the servlet container (in our case, in Tomcat).
Note: Tomcat is a server plugin that
creates an environment for servlet execution. Thus, it is a servlet container or
engine.
The processed results will be sent back to
the invoking HTML/ASP/other web page, which then outputs the results to the
client (browser) in the required
format.
Fig1: System description
What
has been done so far:
·
I have been using parts of JAS
(Java Analysis Studio) to implement the functionality.
·
I extended the parts of JAS that relate to
histogramming and XY-plots to work with database data. So now, it is possible to
extract data dynamically from the database and use it to make basic histograms
and XY plots. These are then output to a web page by the servlets.
·
The user can
specify query parameters via
a form.
·
A skeleton database is in place for present
use.
Ongoing
work:
· Calculating and displaying statistics with the plots.
·
Presentation and arrangement of dynamically
generated graphs on the web page.
On
my agenda:
· Allow more detailed query specification by user.
· Suggest database design to optimize performance.
Possible
problem areas:
· Database design that is different from what would be optimal for the type of data access we require.
· JAS documentation is not detailed. (But Tony Johnson has been very helpful in this regard).
· Manipulating arrangement and presentation of graphs is tedious GUI work.