Database Reference
In-Depth Information
sources that can be analyzed and would benefit from machine learning.
We've moved a long way from just thinking about the web, social media, and
clickstream analytics.
For IT we could take data center management as just one example theme.
Everyserverisarichsourceofdataabouttheenterprise.Beingabletorefine
data center operations presents organizations with fantastic opportunities
for cost savings and streamlined operations. A whole industry has sprung
up around what we call
infrastructure intelligence
. One such company is
Splunk.Theyhavebuiltafantastic productfordelivering insightonourdata
centers and are able to crawl and index vast quantities of machine data to
provide insight on how our data center is being managed.
However, as with all warehouses, this data provides the most value when
we integrate it with other disparate data sources to glean insights and add
value to the business. Otherwise, we simply have another (albeit very cool)
stovepipe application.
Hadoop's Impact on the Data Warehouse Market
This section delves into the Hadoop developer mindset to understand how
the philosophy differs significantly from that of the data warehouse
developer. By understanding the approach taken by Hadoop developers,
we can set ourselves up for a more informed discussion and identify the
most appropriate points of integration between the different environments.
Furthermore, we should also identify opportunities to leverage the
respective technologies for maximum value. It is important to recognize
that both Hadoop and relational databases have distinct advantages and
disadvantages. We have to evaluate the scenario. As architects we are
obligated to leave our zealotry at the door and search for the optimal
solution.
Let's move forward and discuss the following topics:
• Keep everything
• Code first (schema later)
• Model the value
• Throw compute at the problem
