Imhoff: Data warehouse architecture must go beyond the traditionalDate: Jul 10, 2014
The 2014 Pacific Northwest BI Summit takes place in Grants Pass, Oregon, this weekend, bringing together about 20 consultants and vendor executives to discuss business intelligence, analytics and data management trends. Among the participants will be Claudia Imhoff, president of consultancy Intelligent Solutions Inc. and founder of the Boulder BI Brain Trust, a consortium of independent analysts and consultants. In a video interview with SearchDataManagement recorded at last year's summit, Imhoff spoke about an issue that's still front and center for IT, BI and analytics teams: how to expand a traditional data warehouse architecture to incorporate technologies such as Hadoop clusters, data appliances and real-time analytics systems.
The emergence of those technologies and others has "torn apart the architecture that we're so used to, where you have a data warehouse and dependent data marts," Imhoff said in the interview. There's still a need for enterprise data warehouses and the static, historical information typically stored in them, she said. But she added that in the era of big data, cloud computing and mobile applications, data warehouses are no longer enough.
Imhoff and Colin White, founder of consultancy BI Research, have come up with a conceptual design for an extended data warehouse architecture that adds new layers for big data management, exploratory analytics and real-time data analysis. Even with that, though, there's no single path for all organizations to follow. "The big thing to think about is what's the business problem you're trying to solve, and to match the technology to that business problem," Imhoff said in the interview. "Our architecture is a concept. How you physically implement it, that's up to you."
Watch the 10-minute video to hear more of what Imhoff had to say about extending data warehouse architectures to meet today's BI and analytics demands.