The challenge for organizations is figuring out how to get from existing data and application infrastructures to seamlessly ingesting any and all key data types from all sources and rapidly presenting insights to decision makers. A recent survey of 298 data executives and professionals, conducted by UnisphereResearch among members of the Independent Oracle Users Group, finds many organizations lack a cohesive enterprise strategy to capitalize on big data at this time. The largest segment (46%) of data executives in the survey (Big Data, Big Challenges, Big Opportunities: 2012 IOUG Big Data Strategies Survey, September 2012), report they are actually unaware of how big data was being integrated with their business intelligence and analytics systems, while about a third report that their data warehouses are the primary vehicles. That doesn’t mean dispensing existing infrastructures, however. It’s a matter of integrating current capabilities with the new requirements big data brings.
“Organizations are used to managing certain types of data in their RDBMSs, and getting established types of business value from those data types,” says Philip Russom, industry analyst at The Data Warehousing Institute. “Embracing big data asks them to figure out how to manage many new data types, plus get new types of business value from big data. It’s unlikely you’ll get full value from big data from a framework that wasn’t designed for it—the RDBMS that’s at the heart of the average data warehouse. This is not a failure on the part of vendor RDBMSs or users’ designs and architectures for data warehouses.
Just because an elegant luxury sedan isn’t appropriate to off-roading or hauling freight doesn’t mean there’s anything wrong with the sedan or its design.” Current conventional relational database management systems “are not ready, and are ill-prepared to handle big data,” says Wolfgang Seybold, CEO of Cubeware. “Big data requires a profoundly different way of thinking about data, one that can adapt to the concept of collecting and storing structured and unstructured data without a clear idea of how your organization will use it. Big data requires storing and organizing data in a more chaotic way. It’s like the data storage version of Chaos Theory, where you embrace chaos by design.
Open source solutions that handle various aspects of big data management are increasingly finding their way into enterprises as data managers grapple with the challenge. While viable, open source tools often “lack the sophistication that commercial products bring to the table for data analysis, visualization, and presentation that makes data actionable and easy for business use,” cautions Thiru Sampath, a business intelligence architect with X by 2 … Read on.
To access the full version of this article and the complete Best Practices section, go here.