Biomedical Engineering Reference
In-Depth Information
out. To fi ll out the stack it is critical to consider online analytical
processing (OLAP) technologies that support the ability to build cubes.
Cubes are multidimensional data structures that support pre-computed
aggregations and algorithmic slices of data. The only viable option for
open source OLAP databases is Mondrian [31], sponsored by Pentaho,
and it has a broad user base. Mondrian is also used by JasperSoft. For
bringing content from cubes into web-based pivot tables, one should
consider Saiku [32], which has beautiful usability, supports RESTful
OLAP Queries, and a Pentaho plug-in.
20.8 Final architectural overview
This chapter gives an overview of the state of the technology industry
when it comes to scaling clinical analytics using open source software.
Reviewing the architecture in Figure 20.10, we observe that there are
several standard toolsets on the back-end that have risen to become
standards. Mirth, UMLS, and MySQL are mature and capable. Moving
to the right through the fi gure shows products and/or integrations that
are younger but promising. NoSQL and MapReduce are proven by the
giants, but they do require capital, engineering sophistication, and effort
to make use of them. Some of the open source BI tools are mature,
Figure 20.10
Complete analytic architecture
Search WWH ::
Custom Search