Today, you cannot pick up a magazine, read a blog, or hear a webcast without the mention of big data. It has had a profound effect on our traditional analytical architectures, our data management functions, and the analytical outputs.
Analytic environments must keep up with the technological advances and expanding business needs occurring today, but that does not mean chaos will reign. The drivers moving enterprises to extend their data warehouse environments include:
- The need for fast time-to-value to gain business benefits. It is impractical to use traditional data warehouse approaches for all analytical solutions. We must consider extending the environment to include real-time analysis engines, embedded business intelligence (BI), and investigative computing platforms.
- The need for high performance solutions to support new analytic workloads. Uniform data management is no longer viable. Enterprises must match the technologies and costs to business needs and the required analytical workloads.
- The need to change data modeling and integration approaches. Analytical environments need to support new data types, sources, and platforms, as well as new data integration approaches like data blending, data wrangling, schema-on-read, and data repositories.
- The need to change data governance approaches. It is no longer practical to control and govern all forms of data. Enterprises are developing different levels of governance based on security, compliance, quality, and retention needs.
Thoughtful examination of current and future analytical requirements, how to create an expanded architecture, and the development and consistent use of data models and their associated intellectual properties will future-proof the analytical capabilities of an enterprise.
Read the 8-page whitepaper “Ensuring a Sustainable Architecture for Data Analytics” by Claudia Imhoff to explore an analytical architecture that expands on the existing enterprise data warehouse (EDW). It includes additional data sources, storage mechanisms, and data handling techniques needed to support both conventional sources of data and those supplying big data.