The real-time warehouse

The cascade of data flooding into the enterprise is potentially capable of crippling business intelligence (BI) applications, such is the bewildering number of sources of information. And while data warehouses have traditionally been seen as the breakwater that allows businesses to make sense of all the information collected, recently there have been doubts as to whether this approach is responsive enough to meet organisations' requirements.

The size of the challenge is enormous, says Mike Koehler, senior vice president at data warehouse vendor NCR Teradata. The volume of data being generated in a typical business is "doubling every two to three years". Added to that is the wider requirement by users to interrogate that information in a timely fashion.

A recent survey by market watchers Nikkei suggested that three-quarters of businesses are increasing the number of business decisions made daily.

Further out on the horizon, the rise of technologies such as radio frequency identification (RFID) also promises to cause an explosion in the amount of data generated, stored and analysed.

"RFID hasn't yet kicked in to the point where stores are storing the data yet, but as chains such as Wal-Mart start forcing their suppliers to use their data warehouse the market will grow," says Dave Kloc, general manager in Northern Europe for data warehouse appliance vendor Netezza.

Rapid response time

Time sensitivity is a thorny issue for data warehouse vendors. For years, enterprise application makers have been promising to deliver data in real time. But analysing real-time data is tricky: in order to ensure accurate and reliable analysis the data needs to be cleaned before the data warehouse is populated, which introduces latency.

Instead, data warehouse vendors such as Teradata like to promote the concept of ‘right time' or ‘active warehousing', in an effort to emphasise that data only needs to be as fresh as the decisions or business processes require.

The key to cost-effective real-time data warehousing is to avoid the temptation of demanding even fresher data just for the sake of it – data should only be as fresh as its cost and intended use justify. As Stephen Brobst, chief technical officer at Teradata says: "Having data faster than ‘right time' doesn't add value."

For some, the right time is real time. After moving 45 different customer databases to an enterprise data warehouse, staff at Continental Airlines are now able to identify frequent flyers on board their aircrafts, request motorised assistance for delayed passengers with connecting flights and detect fraudulent lost baggage claims.

Using up-to-the-minute data, staff can also run ‘what-if' scenarios to determine the impact of cancellations, delays, or changes to specific flights.

The financial benefits realised by the airline as a result of this approach illustrates just how a real-time response can turn a business around. Continental reports that its $30 million investment in real-time business intelligence has reaped $500 million in increased revenues and cost savings over the last six years.

The extended warehouse

Implementing a successful data warehouse project typically relies on getting any number of competitive vendors to co-operate.

Kim Spenchian, CIO and executive vice president at entertainment industry company Metro-Goldwyn-Mayer (MGM) Studios has seen the problem close up and is well aware of the difficulties.

He asked Teradata and SAP to collaborate on an integrated solution. "I actually had to evangelise the situation to Teradata. I had an uphill battle trying to convince them that they and SAP were not oil and water," he says. "I'm still standing so it hasn't been fatal, but at times it was a little bloody."

The coalescence between data warehouse and enterprise applications has opened up new possibilities for end users. As enterprises extend the reach of their applications to include trusted third parties, pulling buyer and supplier data into the corporate warehouse can add a new level of richness to the information.

At beverage giant Coca Cola, analysing supplier data is a strategic imperative, given that so much of its global operations are outsourced. "We can't be one company and we can't be one big transaction company, but we can be one big data warehouse," says Keith Henry, director of data warehousing at Coca Cola.

The advantage of this approach was charted by analysts when supermarket chain Wal-Mart demanded that its 5,000 suppliers fed data into to its data warehouse. "This made the data warehouse the fulcrum of a lever that would move an entire industry in the direction of just-in-time replenishment and quick response," says Forrester analyst Lou Agosta.

The companies that didn't follow suit, he adds, soon found themselves "out-flanked, out-manoeuvred, and out-of-business" as Wal-Mart became hyper-responsive to dynamic market conditions.


How frequently is your data warehouse updated?
Source: Forrester Research, 2004

Avatar photo

Ben Rossi

Ben was Vitesse Media's editorial director, leading content creation and editorial strategy across all Vitesse products, including its market-leading B2B and consumer magazines, websites, research and...

Related Topics