Although real-time business intelligence (BI) technology isn’t necessarily bleeding edge, it can present a variety
of challenges to organizations that haven’t done their homework before starting a deployment, according to IT and BI analysts.
One potential problem area is that real-time BI isn’t just about the technology, noted Colin White, president of consulting firm BI Research in Ashland, Ore. When you implement a real-time BI system, White said, you might need to think about retraining business users and modifying existing BI and business processes to enable users to take advantage of the faster data access and accelerated decision-making capabilities that real-time BI tools make possible.
“The real question is whether people can deal with reduced latency,” White said. If the answer is no, he added, IT and BI program managers should work to update internal processes to accommodate the real-time BI capabilities – or else think about how real-time their organization really needs to be to deliver the desired business results. In addition, White cautioned that the closer you get to real time, the more robust your IT infrastructure needs to be, both at the BI and analytics end and in the operational systems that are generating the data.
Then there are the tactical issues associated with any new BI project. Often, there is a tendency to try to do too much, too soon, said Lyndsay Wise, president of WiseAnalytics Inc., a Toronto-based consulting firm that focuses on BI and executive dashboard technologies. “Instead of working on one dashboard, many companies want to do everything at once, and then it comes crashing down on them,” she said.
Another key element to consider is project planning, according to Wise. In many organizations, she said, there isn’t enough collaboration between IT, BI and business managers in setting the timelines and deployment plans for projects, including real-time BI initiatives. There also needs to be effective coordination on business needs and the types of BI data required to meet them, Wise added.
On the other hand, Wise is an optimist when it comes to the potential performance impacts of real-time BI technology. “There are basic integration requirements, but a lot of companies will harness extra servers or development servers for the job,” she said. Furthermore, the technical capabilities of BI and data warehousing systems continue to improve, making it easier for companies to support real-time BI. “When I first started consulting in this field, clients required hours every night to back up data and load it into a data warehouse,” Wise said. “Now, that isn’t really much of an issue.”
Real-time BI technology: Already there?
And the underlying technology required for real-time BI is already more ubiquitous than many people might think, said James Kobielus, a senior analyst at Cambridge, Mass.-based Forrester Research Inc. He noted that companies sometimes get tripped up because they don’t realize that real-time capabilities are built into their existing BI tools. For example, they might have in-memory analytics software that could support real-time BI activities, or they might have purchased a complex event processing tool as part of a broader suite of data integration software.
“Many IT pros don’t realize all the software they have available for doing real-time data integration,” Kobielus said. “They don’t have to pay anything additional – they just need to write the data integration scripts, turn [the software] on and then optimize it for real time.”
Of course, it isn’t quite so simple: Companies taking that route need to make sure they have the system bandwidth to support real-time movement of data, Kobielus warned. “They might need to add more capacity in terms of I/O bandwidth and CPUs to get true real-time [throughput], as opposed to just a slightly shorter response time,” he said.
What’s more, in order to manage a combination of batch and real-time BI operations within the same system infrastructure, new tools might be required to monitor processing jobs and dynamically allocate resources to eliminate bottlenecks that are inhibiting real-time BI performance. Typically, real-time BI and data integration technology introduces the need for high-level, policy-based resource provisioning capabilities that can help organizations deliver end-to-end performance optimization, Kobielus said.
In general, he and the other analysts stressed, a successful real-time BI implementation is achievable – as long as organizations are mindful of the real-time approach’s perils and the new demands it can impose on them.
ABOUT THE AUTHOR
Alan R. Earls is a Boston-area freelance writer who focuses on business and technology.