Using MADlib with Greenplum
MAD stands for Magnetic, Agile, and Deep; and lib denotes a library of scalable, par-
allel, and advanced in-database functions. The following figure shows the architecture
of MADlib. The MADlib version used in the following example is v1.1:
Greenplum Database extensions for MADlib would need to be installed on the seg-
ment servers on DCA.
$ pgxn install madlib
$ gppkg -i MADlib
The gppkg utility installs the MADlib extensions on all the Greenplum segment serv-
ers in parallel.
MADlib based in-database analytics is benchmarkedagainst PL/R and is found to be
superior in terms of scalability and performance, and MADlib is a truly parallelized
process as compared to PL/R.