NEW DATABASE TECHNOLOGIES FIT IN
TO THE ENTERPRISE
H ave we entered a new era of database specialization? After many years of relying on one-does-all databases, in which
relational databases, or some variant thereof, were put to use for a variety of jobs—from transaction processing to
departmental databases to web serving—enterprises are entering a new phase of specialization, in which different
enterprise jobs call for specific databases with specific functions.
Core business functions and enterprise transaction processing: This is
where the relational database still rules. Basic enterprise
functions—payments, billing, tracking receipts, deliverables and
payables—are best handled in industrial- class databases built to
interface with all core enterprise applications.
Analytics and business intelligence: When it comes to analytics, the
choice of database environment depends on the type of work being
performed. Much of the unstructured data that comprises many
analytics applications often reside within big data file systems, such
as Hadoop. Thus, there’s an important role for NoSQL databases in
these environments.
IT operational data (logs, systems monitoring): IT professionals need databases they can quickly spin
up while they are in the midst of writing, testing or deploying code. For these purposes, they do not
want to function as accidental DBAs—they need lightweight, quickly displayable databases. In times
gone by, Microsoft Excel spreadsheets or Access databases served this role.
Unstructured data analytics: One of
the hallmarks of big data is the
amount of unstructured, user- and
machine-generated information now
being generated. For the most part,
this data is beyond the reach of the
relational database engines at the
core of enterprises.
Web applications: Web applications generate a lot of session
information, and for this purpose, NoSQL key-value databases may
be the best bet, since they are simple and fast, operating close to
internet time and quickly handling spikes in usage.