LESSON - Performance Matters for Business Intelligence in Financial Services
By Michael Weir, Senior Director of Marketing, ParAccel
Rapidly accumulating business intelligence
data presents an amazing opportunity for all
types of corporations. But “big data” can be
overwhelming—and analyst firm IDC predicts
this data will multiply by 44 times over the
next decade, so it won’t get any easier to
make sense of.
Historical and real-time data could give corporations
a major competitive advantage,
but they must figure out how to access and
analyze it quickly enough to make a difference.
Legacy row-based database solutions
weren’t built for it; they’re slow, expensive
to maintain, and require static and clumsy
work-arounds to get answers. Fortunately,
new analytic database solutions can handle
the avalanche of data by finding smart
and efficient ways to access and analyze it
quickly across many sources.
Financial institutions in particular need
high-performance analytic databases more
than ever. Rapidly changing markets, new
regulations, new investment options, and
new market opportunities must be constantly
and rapidly analyzed for new insights with
risk analysis, which relies on terabytes and
petabytes of historical data. This requires the
superior performance delivered by today’s
analytic databases.
Is Performance Just Query Speed?
Performance is speed sounds simple, but
there are many reasons speed plays an
integral part in the performance of an
analytic database.
First, there is the speed to deploy. After the
market and housing collapse, financial institutions
faced new regulations that required
greater and more intelligent access to their
existing data warehouses. For example, analytic
solutions are now integrated to execute
standardized portfolio stress testing, drilling
into the complexities of a multitude of investments
never previously explored.
Once operational, answers must be accurate
and delivered in time to make a difference.
Financial services firms react immediately
to market turns, predicting them to optimize
trades for their clients and bottom lines. Queries
that identify a trade milliseconds faster
than competitors can mean millions of dollars
in additional revenues.
Performance is speed sounds
simple, but there are many
reasons speed plays an integral
part in the performance of an
analytic database.
In addition to speed, businesses must
be able to administer agile, complex, and
flexible queries. By being iterative and
not incremental when making database
queries, firms can create hundreds of new
query models in the same time it previously
took to develop a few. This leads to deeper
exploration of market change risk analysis,
and entirely new predictions or new business
models. More important, complex
analytic queries against massive amounts of
data can identify fraud risks before fraudulent
transactions are made.
Asking these new questions is critical, but
businesses won’t advance if questions
aren’t answered quickly. With financial firms
trying to predict the future, make instantaneous
trades, or identify fraud before it
happens, queries that previously required
hours or days must be answered in minutes
or seconds.
Speed to return on investment (ROI) is the
final component. The faster these steps
occur—the analytic database is deployed,
queried, and delivers answers—the faster
a company sees ROI. To do so, businesses
should look for analytic databases that reach
these high-performance standards, both in
analytic speed and ease of deployment
and operation.
Meeting the Analytic Challenge
Today’s analytic databases meet the business
intelligence challenge by advancing
performance in novel ways. Combining
shared-nothing massively parallel processing
(MPP) with end-to-end columnar approaches,
overall scale and analytic speed-to-answer are
infinitely improved. Query optimization and
analytic extensibility allow businesses to utilize
and move beyond existing SQL to any level of
complexity without hindering performance.
To quickly realize ROI, software-only analytic
database solutions give corporations the flexibility
and affordability of commodity hardware,
retrieve existing data wherever it’s traditionally
stored, and use the tools and people already
in place. With no tuning or additional workflow
required, DBAs can spend more time running
queries and analyzing answers, and less time
extracting and tuning data.
Whether in finance, retail, or medical
research, businesses rely on insight derived
from their rapidly expanding data stores.
Today’s analytic databases deliver superior
performance by creating smart ways to access
and manage the data explosion across a
broad spectrum of sources, quickly delivering
new business value.
This article originally appeared in the issue of .