While there is considerable buzz around the idea of creating a real-time operation that could consolidate relevant data into a warehouse for analytical slicing, dicing and reporting, users note that considerable obstacles remain.
That was the message of a panel discussion on Monday, "Towards Becoming a Real Time Enterprise", sponsored by application integration maker Ascential Software at the Data Warehousing Institute World Conference.
Several speakers emphasized the subjectivity of the idea of real-time, which is defined differently depending on the company. For instance, Stephen Zander, director of operations and technology, enterprise solutions and services at McKesson, said real-time might vary from an independent firm or a national chain, depending on what their deadlines are. McKesson is a pharmaceuticals distributor that uses several Ascential products, along with SAP's business warehouse software, for such tasks as as ensuring data quality.
Transactions made through McKesson are visible within 24 hours. One of the issues with trying to be real-time is that the company loses the "grace period" to clean up the data to ensure its accuracy, Zander said. At McKesson, IT works with its financial people to explain visibility and to ensure that data fields are properly filled in. Much of the work involved in data warehousing is in cleaning up bad data that might have been put in "10 systems away," he said.
The business case must be there, as well. "A lot of what we're doing is revolving around our ability to move information from our disparate systems and get that into people's hands and make the information really usable in their everyday jobs," said Bob Leo, director of database administration at Landstar System, a transportation service company.
The company uses the Enterprise Integration Suite from Ascential to share data across its Web-based network of independent truckers, agents and customers. "We do things in IT not because we can, but because the business needs us to," Leo said.
The concept of real-time is only applied in those applications where it's really needed, he said. Landstar is more interested in focusing on whatever type of data update is needed -- monthly, nightly, hourly or real-time -- depending on the business case, Leo said.
Although some users request real-time connections, they generally make sense only for a few employees who need to see hourly financials updates, said Danny Siegel, IT director for data management and architecture at pharmaceuticals company Pfizer. The company uses Business Objects Web-based analytical software.
He noted that one of the problems with creating a data warehouse is the need to craft multiple interfaces to it and that every year the cost of maintaining those interfaces rises. "It becomes really big money," he said.