Last update:
Apr 1, 2000
|
|
|
|
Prototyping an object warehouse
|
Takashi Sasaki1,
Hiroko Okazawa2,
Akihiro Shibata1,
Naomi Takashimizu3,
Yoshiyuki Watase1
- Computing Research Center, KEK
- Shizuoka Seika College
- Department of Mathematics and Computer Science,
Shimane University
Presented by:
Takashi Sasaki
We present the design and status of a prototype object warehouse,
an OO version of a data warehouse. Data analysis in HEP is the
process of discovering meaningful new correlations, patterns and
trends by scanning through large amounts of stored data, using
statistical and mathematical techniques as well as pattern recognition
technologies. This method is called as "data mining" in the
information science. A big information center specialized for fast
data mining, a data warehouse, has been implemented in many
organizations. We have fully taken knowledge in these area into
account designing the system.
The key issue in data mining is the performance of databases
accessed from many clients. Objectivity/DB, which is widely used
in HEP, has a bottle neck at the I/O server, AMS. We improved the
situation using a SAN (Storage Area Network) technology to share
a disk directory among clients over the fiber channel. We will
discuss the performance of Objectivity/DB on the protype object
warehouse.
We are also developing software tools for an object warehouse based
on already existing data mining techniques for fast and efficient
data analysis in HEP.
Short Paper:  |
|
Bulletins
| Home | Bulletins | Committees | Scientific Program | Docs by topics | Social Event | Conference Location | Secretariat |
Privacy Policy |