Computer Technology Review just posted a piece we wrote on Scaling Business Analytics. This has become a hot topic recently and 2007 was marked by three major acquisitions in this space...SAP bought Business Objects for almost $7 billion, IBM bought Cognos for around $5 billion, and Oracle bought Hyperion earlier in the year for over $3 billion.
$15 billion to up the ante in the business intelligence (BI) arena will certainly make 2008 interesting.
So the race is on to who can serve up the best BI solutions to the greatest number of users...empowering decision making throughout the corporation.
In a recent piece titled, Forecasting the Future of Business Intelligence, Ephraim Schwartz writes the about the reasons businesses are increasing BI use:
...it may have taken five or 10 years for the
enterprise to understand the significance of BI, but once BI was
adopted, the learning curve proved short, and now that enterprises have
seen the light, they want more. The enterprise does not only want the
ability to look back; it wants the ability to look forward.
...For example, companies use forecasting for
predicting inventory demand, building customized analytical
applications. In financial services, forecasting can optimize
portfolios for higher performance.
We're also seeing a greater need for real-time analysis. Jessica Twentyman wrote this piece in ComputerWeekly.com discussing ways in which administrators have dealt with this issue in the past...primarily moving data around between transactional and analytical systems.
Real-time business intelligence
In
most finance departments, data is moved from transaction systems to a
"staging post" before it is analysed. But for some, this can create
some problems when it comes to month-end reporting. Journal entries may
take some time to be moved from operational to analytical data stores,
and that can make reconciling ledger accounts challenging.
"Although
this latency is generally acceptable for trend analysis and
forecasting, traditional datawarehouses cannot keep pace with today's
business intelligence requirements for fast and accurate data," says
Louella Fernandes, principal analyst at research firm Quocirca. They
were not designed to deliver complex analytics on terabytes of data
quickly, and as the volume of data used in organisations grows,
extracting information becomes more time-consuming and complex.
Moving data around, and replicating data within an enterprise is a complex way to get around multiple requirements for how data is captured, stored, processed, and analyzed.
But increasingly we see opportunities to let the data remain within a simplified, single tier of storage, and provide the access for analytics through a caching appliance.
With data accessible through a caching appliance the ability to serve multiple departments simultaneously becomes significantly easier as the appliance delivers plenty of IOPS, low-latency, and throughput without impacting the source database. The need to replicate data for performance goes away, and the response time for data coming from a caching appliance can be orders of magnitude less than from traditional disk-based storage systems.
All of these factors combine to make scaling analytics applications far easier, and are likely to feed the thirst for more comprehensive business analytics across the board.
