A business intelligence data warehouse is a key component for a successful BI implementation. While software can provide some of the power needed to get data and information into reports and other forms, a good underlying data source is the workhorse. Garbage in, garbage out is definitely an appropriate idea to keep in mind with this concept.
While most systems can report against transactional, operational, or other near real-time data sources, most business intelligence experts agree that a traditional datawarehouse is the ideal situation. The advantage to a dedicated source is that it will be optimized for reporting and intelligence, not transaction or real-time activity. It also centralizes and consolidates data sources and prevents silos of information, while making standards and ensuring security across an organization. This is the model that most BI vendors would prefer an organization use. While ideal, many organizations can’t always get this in place, or departmental data can’t be incorporated to the big data scheme.
This traditional notion is starting to get redefined, with new ways of computing though. One of these new concepts is the idea of a logical model, which can source from multiple sources and uses an open source shared server technology like Hadoop to manage and get access to the data.
Volumes could be written about the theories, systems, and technical discussions about databases and data storage. A good designer should be able to architect a structure that can support the data that you want now and in the future. It’s a common mistake for organizations to box themselves in with a poorly designed or constructed data warehouse. The issues can lie in the software or hardware too. Particularly with the number of mergers occurring in this space, sometimes lines can be discontinued. IBM buying Netezza and HP buying Vertica are both interesting and are very related to the Big Data issues surrounding BI.
The BI tool that you use should be able to work with your current or proposed data structure. You need to make sure it will scale properly and give the flexibility to fill in the gaps that the transaction, ETL, or DW processes can’t or won’t cover.
Copyright 2011-2019 BusinessIntelligenceBase.com