Data Warehouse options for Analytics

Increased competition and a fast moving
business environment means that there is
a much greater demand for analytics and
business intelligence than ever before, to
better understand customers for example.
This urgency is frequently combined with a
desire not just to analyse the past but also to
help to predict the future: for instance, how a
customer is likely to react to a special offer.
Moreover, companies increasingly want to
do this sort of analysis in as close to real-
time as possible, because environments like
call centres require information that is both
up-to-date and immediate. Moreover, the
audience scope for analytics has also greatly
expanded. Historically, business intelligence
was limited to a few expert business
analysts and the production of standard
reports. Today, however, you may want to
provide wider access to a broad business
or customer community for ad hoc queries,
to embed queries into business processes
or operational environments such as call
centres, or present information within a real-
time dashboard.

All of this change is further exacerbated by
the growth in the amount of data that needs
to be analysed to meet these needs. Many
data warehouses, especially those supporting
critical business processes, must now expect
to have to incorporate tens or even hundreds
of terabytes of raw data. Further, the variety of
data that needs to be analyzed is also evolving,
producing a need for techniques to query non-
traditional types of data such as text, video
images, geospatial data and so forth.

Given this hugely more complex landscape
for queries and analytics it is perhaps
no surprise that in recent years the data
warehousing market (which serves these
business intelligence needs) has itself become
more complicated. Organisations now have
a number of new data warehousing options
offering a variety of approaches that have
changed the data warehousing landscape.

It should be clear that different organisations
will have different requirements when
it comes to business intelligence and
analytics. However, the difficulty that arises
is the question of how to match that set
of requirements against the various data
warehouse technologies and products that
are available. Each of these options is better
suited to some environments than others.

In this paper we will start by considering
some of the most important generic issues
that need to be considered when selecting the
provider of a data warehouse. We will then
consider these data warehouse options in the
light of some specific business applications
for analytics, including on-demand analytics
and advanced analytics. While these analytic
uses reflect specific applications common
to customers of Sybase IQ analytics server,
Sybase’s data warehouse product, we will
consider the relative merits of various
vendors’ technologies or types of technology,
throughout the discussions that follow.
Vendors that will be considered include
Sybase IQ, a column-based analytics server;
traditional row-based databases from Oracle,
Microsoft SQL Server, IBM DB2 and MySQL;
Teradata; and the new data warehouse
appliances from Netezza, Hewlett-Packard,
Dataupia and DATAllegro.

Free Download (subject to terms)