Database Reference
In-Depth Information
Component Integration Within
the IBM Big Data Platform
All of the components of the IBM Big Data platform are well integrated. In
this section, we briefly highlight some of this integration, and provide point-
ers to the chapters where you can find more detailed information.
InfoSphere BigInsights
BigInsights is deeply integrated throughout the IBM Big Data platform.
BigInsights can accept real-time input from Streams, such as insights or deci-
sions made by analytic algorithms deployed to the Streams run time, and
then perform additional analysis on a deeper data set (for example, to test
and remodel the algorithm for promotion back to Streams).
BigInsights is also integrated with the analytics-based IBM PureData Sys-
tems and other IBM data warehouse solutions, such as InfoSphere Warehouse.
This enables BigInsights to write data to a warehouse and to accept input
from it too. It's also possible to access data in BigInsights by using SQL.
In-database functions that present SQL-wrapped user-defined functions
enable you to call a BigInsights job from within the database. (Note, however,
that the laws of physics apply here: You're not going to leverage these func-
tions to pull trillions of rows across the network into your warehouse, you
more likely want to invoke a job and retrieve a scored result.) This integration
point is helpful when BigInsights is used to analyze raw data or to perform
experimental analysis, and then share insight with the data warehouse.
There's integration between BigInsights and InfoSphere Information Server
(Information Server). Information Server can provide information to Hadoop
or take data from Hadoop and integrate it to any number of other systems.
BigInsights is also integrated with Guardium to provide real-time system
monitoring and security. InfoSphere Optim Masking Solution can mask
BigInsights data to ensure that it's protected and secure. And InfoSphere
Optim can store archived data in BigInsights, making it available for analysis.
IBM Research has also developed integration points between MDM and
BigInsights, including the ability to analyze raw data and to determine which
master entities are to be loaded into MDM. You can find more details about
this integration in Chapters 5 through 8.
