Transcript Churches_DB_usage

Database Developments
David Churches, Cardiff University
GEO Data Analysis Workshop
AEI
6 April, 2016
Introduction and terminology
Á
A database is meant here to mean a
collection of tables
Á
A table is a collection of data arranged into
rows and columns
Á
A database management system is some
piece of software which the user interacts
with to get at the tables e.g. mySQL, DB2 but this is also usually called the database
Database Usage so far
Á For our work on GEO++ and the inspiral
search code, we have been using mySQL
databases set up on the following machines:
The Hanover Beowulf
The Cardiff mini Beowulf
Each of our own laptops
Á In each case, the databases and tables are
mirrors of each other
A Specific Example (I)
Inspiral Search Code
(II)
Parallel MPI C Code
Using C functions to interact with
database
(I)
(III)
process table
sngl_inspiral
table
program name
program name
start time
start time
username
process_id
unix process id
unique process id
process id
amplitude of event
time of event
masses
chirp times
A Specific Example (II)
Excess Power Code
(II)
Parallel MPI C++ Code, part of GEO++
Using C ++ functions to interact with
database
(I)
process table
program name
program name
start time
start time
username
process_id
unix process id
unique process id
(III)
sngl_burst table
process id
amplitude of event
time of event
central frequency
bandwidth
Database tables in use
process
sngl_inspiral
sngl_burst
Tools being used for database
interactions
Á The mySQL command line
;
Simple to use but rather tedious
Á The mySQL GUI
;
Better, but you cannot do any further numerical/graphical
analysis upon the results
Á Triana
;
Can do everything the mySQL GUI can, but can also perform
subsequent data analysis. However, it is a bit slower at
tabulating the results.
Á Matlab
;
Soma......
Open Issues / discussion
Á Should we use DB2?
;
;
;
We have a copy and so we can try it out
Are there implications for interaction with LIGO?
Open source mySQL vs commercial DB2 ?
Á At the moment we cannot access the
mySQL server on the Hanover beowulf
because of the firewall
;
;
This means that we are copying the files which contain the
mySQL data from Hanover to our own machines
Should be simple to fix, however
Á The continuous waves group have not been
using any database