Database Reference
In-Depth Information
brand has become synonymous with Hadoop, largely through the reach of
their Cloudera Distribution for Hadoop (CDH). Although BigInsights
includes the IBM Distribution for Hadoop, extending support to CDH does
not reflect a contradictory approach. Rather, it's another way in which IBM
supports a strong and unified Hadoop community. Because it's all based on
open source code, the IBM Distribution for Hadoop and CDH are really not
too different.
The main purpose behind IBM's support for Cloudera (and eventually,
other open source Hadoop distributions) is to emphasize the value of the IBM
Big Data platform. In short, even if you're not using the IBM Distribution for
Hadoop, you can take advantage of the advanced analytics, integration, and
tooling capabilities in BigInsights. Clients obtain CDH from Cloudera, install
their CDH cluster, and then BigInsights enterprise features can be layered on
top of CDH. For CDH users, this is a great story: they no longer have to make
a difficult choice to leave the comfort of their installation to use BigInsights.
The point here is that the choice of which distribution you use is less impor-
tant than your approach to analytics higher in the stack.
Analytics: Exploration,
Development, and Deployment
The biggest challenge facing organizations that have adopted Hadoop is
how to quickly and easily derive value from the data that's stored in their
clusters. There are many factors to consider here, some of which we've touched
on earlier. For example: the difficulty of parallelizing many analytics algorithms;
making data stored in Hadoop accessible for business analysts and data
scientists; and dealing with really messy data. BigInsights addresses all of
these issues by applying an integrated platform-based approach across its
various components.
As an analytics platform for Hadoop, BigInsights supports three classes of
analytic users: line-of-business analysts, data scientists, and application devel-
opers. In addition, the analytics and development components of BigInsights
feature lifecycle management tooling, which enables developers to easily deploy
analytic applications to line-of-business users and data scientists. Figure 5-2
shows all of the analytic components in BigInsights. We describe each of
these components in this section. We also cover the application lifecycle
