Databases Reference
In-Depth Information
Putting It All Together
Earlier in this chapter, we described the components in a complete end-to-end analytics
architecture. Now, we'll take a look at what an end-to-end analytics infrastructure might
look like with Oracle components for business intelligence, data warehousing, and Big
Data. Then we'll discuss some best practices.
A Complete Analytics Infrastructure
How extensive an analytics infrastructure your organization needs depends on your
business needs. At a minimum, most organizations have a designated data store/data
warehouse containing extracted data from key sources and a frontend BI tool or re‐
porting engine. A more extensive footprint might also include Big Data, a variety of
ETL, data movement, and event processing solutions (such as Oracle Event Processing,
or OEP), data quality and master data management solutions, a mixture of reporting,
dashboard, and ad hoc query tools, advanced analytics tools for statistical analysis and
data mining, and a real-time recommendation engine.
A complex analytics infrastructure that contains Oracle components might look like
the following diagram in
Figure 10-6
.
Figure 10-6. Oracle Analytics Infrastructure Footprint
In this scenario, we find structured data in the Oracle data warehouse gathered from
transactional data sources. A variety of other data sources, including sensor data, social
media feeds, and web log information, are analyzed in a Hadoop cluster. We use a data
discovery tool, Endeca, to explore all of our data, structured and unstructured. We
populate the data warehouse with MapReduce output containing the data of value we
