Data warehousing
The general idea of data warehousing was invented in 1980s, as a response to growing importance of data flow. Higher and higher information volumes have challenged the methods of accessing data, which - finally - haven't passed the test. New, more efficient solutions have become a necessity. For the first time, name 'data warehousing' was used in 1988, by IBM researchers and only two years later the first database management system for data warehousing was introduced. In a few months, further solutions appeared, but no one really expected such enormously fast growth of popularity and importance of data warehousing.
What were its reasons?
Strengths and weaknesses of implementing a Data warehouse
Data warehousing wouldn't reach its popularity without a few significant features it offers:
- Information management, due to extremely high importance of information and complex business strategies, demands often a quick access to data from different sources - different departments, different tasks, different models. Data warehousing facilitates the access to information, offering one data model for every source. A common system of data storing results in easier analysis and reporting.
- As long as analysis and reporting has the most important meaning in performance management, data consistence requires the highest care. While loading data into DW, all the inconsistencies are immediately discovered and, then, fixed.
- Almost every piece of information can be used many times, from different points of view. Therefore, it's important to store data even longer than it's been already stated, longer than typical source systems allow. With data warehouses, the time data is stored, depends only on customers' needs.
- Even the best computers and servers have their efficiency limits - the more tasks or the more memory-consuming task, the longer time is needed for each. Data warehouses do not base on operational systems, therefore working on them doesn't cause significant slowing down.
- Having two or more data warehouses doesn't mean a necessity to choose one of them - they might really well work together, mutually supplying their features while, for example, using business applications.
- Data warehouses efficiently support decision making, due to weight attached to report applications.
Even the best solution never is perfect. Data warehouses have also a few disadvantages.
- While data warehouses work perfectly with structured data, the solutions responsible for unstructured one might be better.
- Fulfilling warehouse with data requires extracting, transforming and loading - processes inevitably causing latency.
- Data warehouses' maintenance generates usually significant costs.
- Data warehouse's validity passes relatively quickly.
- Data warehouses are often similar to operational systems and multiplying the same functionality generates superfluous costs that might have been easily omitted.
There are two sides to every story and so is to data warehousing. Significant benefits might overshadow disadvantages but obviously that's not a rule. Therefore, grows the importance of choosing best data warehousing platform. According to variety of vendors, there are more and more significant differences among each platforms. After all, a few questions become inalienable.