Simplify to Overcome Historical IT Complexity
Data virtualization's simplified information access and faster time-to-solution is especially useful as an enabler for more agile analytics and BI
Is Data Virtualization the Fast Path to BI Agility? describes how the architectures of most business intelligence systems are based on a complex chain of data stores starting with production databases, data staging areas, a data warehouse, dependent data marts, and personal data stores. Simply maintaining this complexity is overwhelming IT today.
These classic BI architectures served business well for the last twenty years. However, considering the need for more agility, they have some disadvantages:
- Duplication of data
- Non-shared meta data specifications
- Limited flexibility
- Decrease of data quality
- Limited support for operational reporting:
- Limited support for reporting on unstructured and external data"
From a different point of view, SOA World's Zettabytes of Data and Beyond describes the challenges of force-fitting development methods that were appropriate for earlier times when less data complexity was the norm.
In addition, the proliferation of fit-for-purpose data stores including data warehouse appliances, Hadoop-based file systems, and a range of No-SQL data stores are breaking the hegemony of the traditional data warehouse as the "best" solution to the enterprise-level data integration problem. The business and IT impact of these new approaches can be explored in the Virtualization Magazine article NoSQL and Data Virtualization - Soon to Be Best Friends.
Self-Service Analytics and BI are Important Too!
Responding to constantly changing business demands for analytics and BI is a daunting effort.
Mergers and acquisitions and evolving supply chains require new comparisons and aggregations. The explosion of social media drives demand for new customer insights. Mobile computing changes form factors. And self-service BI puts users in the driver's seat.
Business Taking Charge of Analytics and BI
In true Darwinian fashion, the business side of most organizations is now taking greater responsibility for fulfilling its own information needs rather than depending solely on already-burdened IT resources.
For example, in a 2011 survey of over 625 business and IT professionals entitled Self-Service Business Intelligence: TDWI Best Practices Report, @TDWI July 2011,The Data Warehousing Institute (TDWI) identified the following top five factors driving businesses toward self-service business intelligence:
- Constantly changing business needs (65%)
- IT's inability to satisfy new requests in a timely manner (57%)
- The need to be a more analytics-driven organization (54%)
- Slow and untimely access to information (47%)
- Business user dissatisfaction with IT-delivered BI capabilities (34%)
In the same survey report, authors Claudia Imhoff and Colin White suggest that IT's focus shifts toward making it easier for business users "to access the growing number of dispersed data sources that exist in most organizations."
Examples Imhoff and White cite include:
- providing friendlier business views of source data
- improving on-demand access to data across multiple data sources
- enabling data discovery and search functions
- supporting access to other types of data, such as unstructured documents; and more.
Data Virtualization to the Self-Service Rescue
In the TDWI survey, 60% of respondents rated business views of source data as "very important," and 44% said on-demand access to multiple data sources using data federation technologies was "very important."
According to Imhoff and White, "Data virtualization and associated data federation technologies enable BI/DW builders to build shared business views of multiple data sources so that the users do not have to be concerned about the physical location or structure of the data.
These views are sometimes known as virtual business views because, from an application perspective, the data appears to be consolidated in a single logical data store. In fact, it may be managed in multiple physical data structures on several different servers.
Data virtualization platforms such as the Composite Data Virtualization Platform support access to different types of data sources, including relational databases, non-relational systems, application package databases, flat files, Web data feeds, and Web services.